Sediment Modeling Panel (part 4): Forecasting With Sediment Models and Q&a

Most morphological model developers are also morphological modelers. Their model development reflects their own modeling experiences, including recurring problems and the best practices they’ve developed over time. This panel discussion – the final conversation in a four-part session – explores the final phase of most modeling studies, forecasting and alternative evaluation.

One primarily issue related to forecasting or predicting with morphologic models is quantifying the uncertainty in this prediction. Two basic forms of this uncertainty are epistemic (knowledge and measurement gaps) and aleatory (natural variability). The epistemic uncertainty includes limitations in the model formulations, uncertainty in model parameters, uncertainty in initial or boundary conditions, and errors introduced by the numerical method. Aleatory errors include such unknowns as the future hydrology or future sediment loading. Future hydrology is one of the most important modeling decisions that well-calibrated sediment models face in the forecasting stage. Stationarity issues can also complicate the move from calibration (reproducing the past) to forecasting future processes.

In this session, the current or former sediment transport leads for ADH, SRH-1D, and HEC-RAS 1D and 2D sediment development, will discuss how the develop their future-condition hydrology and sediment parameters.

This session is likely to cover topics such as: • Selecting a Future Flow Series. • Dealing With Non-Stationarity when Forecasting. • Aleatory vs Epistemic Uncertainty • Sensitivity Analysis and Stochastic Simulations. • Methods to quantify prediction uncertainty • Describing model uncertainty

This session will also reserve extended time for audience questions. If you have ever wanted to ask four morphological sediment modelers a public question, this will be the forum.