Prediction of wave runup on beaches using Gene-Expression Programming and empirical relationships

摘要: Abstract This paper assesses the accuracy of seven empirical models and an explicit Gene-Expression Programming (GEP) model to predict wave runup against a large dataset observations. Observations consist field laboratory measurements include wide array beach types with varying sediment sizes (from fine sand cobbles) bed roughness smooth steel asphalt). We show that best performing in literature are prone significant errors (minimum RMSE 1.05 m NMSE 0.23) when used unseen data, i.e., uncalibrated models; however, overall error values correlations significantly reduced optimised for dataset. The use Hunt type scaling additional parameter induced setup. predictive ability GEP model, which better captures complex nonlinear effects key factors on length, resulted statistically improvement capacity comparison all other assessed here, even data. Wave height, wavelength, slope shown be three primary influencing runup, grain size/bed having smaller, but still influence runup. r2 existing (which takes form Holman (1986) Atkinson et al. (2017) their M2 model) was 0.77, 0.85 m. These were improved 0.82 (6% increase) 0.75 m (12% decrease) GEP-based model. sensitivity proposed each input variable is via partial derivative analysis. results demonstrate higher small steepness