Evaluation of Parameter and Model Uncertainty in Simple Applications of a 1D Sediment Transport Model

Abstract

This paper separately evaluates two methods from Bayesian Statistics to estimate parameter and model uncertainty in simulations from a one-dimensional (1D) sediment transport model. The first method, multivariate shuffled complex evolution metropolis-uncertainty analysis (MSU), is an algorithm that identifies the most likely parameter values and estimates parameter uncertainty for models with multiple outputs. The second method, Bayesian model averaging (BMA), determines a combined prediction based on three sediment transport equations that are calibrated with MSU and evaluates the uncertainty associated with the selection of the transport equation. These tools are applied to simulations of three flume experiments. For these cases, MSU does not converge substantially faster than a previously used and simpler parameter uncertainty method, but its ability to consider correlation between parameters improves its estimate of the uncertainty. Also, the BMA results suggest that a combination of transport equations usually provides a better forecast than using an individual equation, and the selection of a single transport equation substantially increases the overall uncertainty in the model forecasts.