Sensitivity and predictive capability studies on a finite element surge simulation model

This paper reports the sensitivity and predictive capability studies carrie d out on the finite element model for storm surge simulation along the east coast of India. In this model, an explicit finite element scheme with 9-no ded Lagrangian quadratic isoparametric elements was used to discretize the study area and the governing equations. The integration of dependent variab les over an element was performed using two-dimensional Simpson's rule, whi ch leads to uncoupling of the system of equations and yields an explicit sc heme for determining the flow parameters evolving with time. In the present work, the model has been benchmarked using an International software MIKE2 1. Sensitivity studies are carried out to evaluate the influence of the len gth of the ocean shelf and its depth at the coast in predicting the storm s urges along the coast. The results show that the model is not sensitive to these parameters. Then the model is calibrated by simulating the storm surg es induced by the 1977- Andhra cyclone, which is one of the severe cyclones in recent years along the east coast of India. The predictive capability o f the calibrated model is assessed by simulating the surges for a few past cyclonic storms, which have crossed the east coast of India.