Multi Station Calibration of 3D Flexible Mesh Model: A Case Study of the Columbia Estuary☆

Abstract A model calibration based on the distributed multi stations approach is necessary towards model implementation in the operational phase. In this study, a tree dimensional (3D) hydrodynamic and salinity dynamic model of an estuary was simulated using DFlow Flexible Mesh program, which is developed by Deltares. Specifically, this research was focused on the Columbia Estuary case study, which is situated in Oregon, United States. The preconfigured model was calibrated based on 15 measurement stations that are spread along the estuary. Furthermore, a detail portion data with an average interval of 1 minute were used during the calibration process. The model performances were improved by considering the data denial concept. The data denial concept was introduced by neglecting inconsistence data across its temporal and spatial variability. In this particular case, it was revealed that the downstream data, which have high salinity value, tends to produce high contribution to the root mean square error of the model result. In conclusion, the upstream data have immense variable fluctuation rate and therefore it is more sensitive to give lower coefficient of determination. Therefore, there must be a trade of between good estuary model performance and upstream station data reliability.