UNCERTAINTY IN THE RELATIONSHIP BETWEEN DISCHARGE AND PARAMETERS IN MODELS OF LONGITUDINAL POLLUTANT TRANSPORT

When calibrating a pollutant transport model for a River, a tracer study is usually performed under steady-state flow conditions. However, the model parameters, estimated using the tracer data, depend on the river discharge. This paper investigates the uncertainty of the relationship between discharge and the parameters that describe the transport of conservative pollutants using transient storage and aggregated dead zone models. We apply a Bayesian statistical approach to derive the cumulative distribution of a range of model parameters conditioned on discharge. The set of tracer data consists of eighteen tracer concentration profiles taken at two cross-sections from the Murray Burn, a stream flowing through the Heriot-Watt University Campus at Riccarton in Edinburgh. For both models the uncertainty of the estimated parameters increases with increasing discharge. After the calibration, both models' parameters were parameterised as an exponential function of flow. The newly developed parameterised models were once again calibrated using the Generalised Likelihood Uncertainty Estimation (GLUE) technique adjusted for the case of the heteroscedastic variance and validated against the tracer experiments data not used in the calibration stage. In validation runs, both models gave similarly good predictions of concentration profiles.