1. Application of hierarchical Bayesian unmixing models in river sediment source apportionment
- Author
-
Blake, William H., Smith, Hugh G., Navas Izquierdo, Ana, Bodé, Samuel, Goddard, Rupert, Kuzyk, Zou Zou A., Lennard, Amy T., Lobb, David A., Owens, Philip N., Palazón Tabuenca, Leticia, Petticrew, Ellen L., Gaspar Ferrer, Leticia, Brian Stock (7), Blake, William H., Smith, Hugh G., Navas Izquierdo, Ana, Bodé, Samuel, Goddard, Rupert, Kuzyk, Zou Zou A., Lennard, Amy T., Lobb, David A., Owens, Philip N., Palazón Tabuenca, Leticia, Petticrew, Ellen L., Gaspar Ferrer, Leticia, and Brian Stock (7)
- Abstract
In this contribution, key components of the model are evaluated in the context of common experimental designs for sediment fingerprinting studies namely simple, nested and distributed catchment sampling programmes. Illustrative examples using geochemical and compound specific stable isotope datasets are presented and used to discuss best practice with specific attention to (1) the tracer selection process, (2) incorporation of fixed effects relating to sample timeframe and sediment type in the modelling process, (3) deriving and using informative priors in sediment fingerprinting context and (4) transparency of the process and replication of model results by other users.
- Published
- 2016