Bayesian Sediment Transport Model for Uni-size Bed Load
Location
Eccles Conference Center
Event Website
http://water.usu.edu/
Start Date
3-29-2011 9:20 AM
End Date
3-29-2011 9:40 AM
Description
Recent studies in the field of sediment transport in fluvial systems underscore the difficulty in reliably estimating transport model parameters, collecting accurate observations, and making predictions due to measurement error and uncertainty in conceptual models. While much of the initial research in sediment transport recognized the process as being inherently statistical, developing a sediment transport model in a statistical framework that can manage and account for measurement error, process uncertainty, and provide credible intervals of parameter estimates and transport predictions still presents many opportunities for discovery. In this research, we propose a uni-size bed load sediment transport model in which inference on the critical Shields Number and measurement error along with credible intervals of transport prediction are sought. This model is realized through the application of a Bayesian framework in which posterior distributions of model parameters are evaluated given both simulated data and uni-size sediment transport observations pulled from the literature. Our model provides a sediment rating curve that is delineated in terms of credible intervals for prediction. Credible ranges of the critical Shields Number and process/measurement uncertainty are inferred laboratory flume data, the proposed framework can be applied to fluvial systems of arbitrary size. Additionally, our model presents opportunities to evaluated the suitability of different transport relations. The initial findings of applying a sediment transport relation in a Bayesian framework show great promise for management uncertainty and variability in associated process models and measurements.
Bayesian Sediment Transport Model for Uni-size Bed Load
Eccles Conference Center
Recent studies in the field of sediment transport in fluvial systems underscore the difficulty in reliably estimating transport model parameters, collecting accurate observations, and making predictions due to measurement error and uncertainty in conceptual models. While much of the initial research in sediment transport recognized the process as being inherently statistical, developing a sediment transport model in a statistical framework that can manage and account for measurement error, process uncertainty, and provide credible intervals of parameter estimates and transport predictions still presents many opportunities for discovery. In this research, we propose a uni-size bed load sediment transport model in which inference on the critical Shields Number and measurement error along with credible intervals of transport prediction are sought. This model is realized through the application of a Bayesian framework in which posterior distributions of model parameters are evaluated given both simulated data and uni-size sediment transport observations pulled from the literature. Our model provides a sediment rating curve that is delineated in terms of credible intervals for prediction. Credible ranges of the critical Shields Number and process/measurement uncertainty are inferred laboratory flume data, the proposed framework can be applied to fluvial systems of arbitrary size. Additionally, our model presents opportunities to evaluated the suitability of different transport relations. The initial findings of applying a sediment transport relation in a Bayesian framework show great promise for management uncertainty and variability in associated process models and measurements.
https://digitalcommons.usu.edu/runoff/2011/AllAbstracts/9