Local pier scour remains one of the leading causes of bridge failure, calling for predictions that are both accurate and uncertainty-aware. This study develops an interpretable data-driven framework that couples CatBoost (Categorial Gradient Boosting) for deterministic point prediction with NGBoost (Natural Gradient Boosting) for probabilistic prediction. Both models are trained on a laboratory dataset of 552 measurements of local scour at bridge piers using non-dimensional inputs (y/b, V/Vc, b/d50, Fr). Model performance was quantitatively evaluated using standard regression metrics, and interpretability was provided through SHAP (Shapley Additive Explanations) analysis. Monte Carlo–based reliability analysis linked the predicted scour depths to a reliability index β and exceedance probability through a simple multiplicative correction factor. On the held-out test set, CatBoost offers slightly higher point-prediction accuracy, while NGBoost yields well-calibrated prediction intervals with empirical coverages close to the nominal 68% and 95% levels. This framework delivers accurate, interpretable, and uncertainty-aware scour estimates for target-reliability, risk-informed bridge design.
31 December 2025 in Journal of the Korean Society of Hazard Mitigation
The interactions of river flows with bridge piers generate strong vortices and locally accelerated velocities that disturb the bed and produce local scour. In this regard, approaches for predicting scour depth have evolved from empirical formulas derived from field and laboratory data to numerical models that solve the governing hydrodynamics and, more recently, machine-learning methods that learn patterns from data. Herein, we propose a model for predicting scour depth using NGBoost, which predicts full probability distributions rather than single values, thereby quantifying the predictive uncertainty and allowing for safety-oriented decisions through predictive intervals. To interpret the model, we apply Shapley additive explanations to identify the hydraulic and geomorphological variables that have the greatest influence on the predictions. When compared with representative empirical equations, the NGBoost model shows better accuracy, indicating enhanced performance for pier-scour prediction. Overall, NGBoost offers an uncertainty-aware interpretable framework to obtain a more reliable assessment of scour risk around bridge piers.