Hierarchical Bayesian modeling of root distribution for the geotechnical analysis of bioengineered systems
Hierarchical Bayesian modeling of root distribution for the geotechnical analysis of bioengineered systems
The geotechnical modeling of rooted soils is challenging due to the strong spatial variability of root systems and to the epistemic uncertainties stemming from the typically limited and sparse available datasets and from the use of simplified analytical models for the estimation of root-induced cohesion. This study introduces a flexible Hierarchical Bayesian Regression model for the modeling of the vertical spatial variability of root area ratio. The proposed approach includes a non-monotonic functional form which ensures improved coherency with empirical data. The model is trained on a preliminary multi-species dataset and subsequently validated on new data, demonstrating the utility of the Bayesian approach in reducing epistemic uncertainty in typical scenarios of limited evidence. The probabilistic outputs of the approach can be used directly in reliability-based design of bioengineered geotechnical systems.
Stefano Collico; Marco Uzielli
3rd International Workshop on Soil-Vegetation-Atmosphere Interaction (RootS2025)
2b. Field scale characterisation of the bio-hydro-chemo-mechanical behaviour of rooted soils
https://doi.org/10.53243/RootS2025-22
Contact Information
- Secretary General - Dr A M McNamara
- Email: [email protected]
- Phone: +44 20 7040 8154
- City, University of London
- Northampton Square
- London EC1V 0HB. United Kingdom