Multivariate Adaptive Regression Splines model for sediment incipient deposition in urban drainage channels

It is already found that channel cross-section shape impacts on sediment transport condition in open channels. To this end, this study is aimed to generalize the incipient deposition models through incorporating a cross-section shape factor into the models. Multivariate Adaptive Regression Splines (MARS) technique is used for the modelling and results are compared with the developed model based on conventional Multi Non-Linear Regression (MNLR) method. Comparison of developed models in this study with those existing in the literature indicates that cross-section specific models may have poor performances on channels with variety of cross-sections. MARS and MNLR models as general incipient deposition models outperform cross-section specific models which can be linked to the incorporating of shape factor as input of the models. Explicit equations are suggested which can be used as practical tools for stormwater urban derange channel design.

Keywords:

Cross-section shape , Multivariate Adaptive Regression Splines , Sediment transport , Urban drainage