Accurate predictions of storm-tide are of vital importance for many coastal areas. Along North Sea coasts, reliable storm-tide predictions are of crucial importance as a large portion of the coastal zones is not only below mean sea level but also characterized by frequent storms. Currently, the nature of mutual nonlinear interactions between the extreme storm-tide cannot yet be solved by conventional hydrodynamic/statistical models alone. Alternatively, a pragmatic data-driven approach is proposed and implemented in Cuxhaven and Sylt, which combines the predictions of hydrodynamic model (TELEMAC) and nonlinear autoregressive with exogenous input (NARX) recurrent neural networks model. This approach, called hybrid TELEMAC-NARX model, can implicitly account for the nonlinear interactions among all storm-tide components. This enables the substantial errors in both magnitude and timing of the results predicted by the hydrodynamic model to be corrected. The maximum effect of the nonlinear interaction at Cuxhaven and Sylt reach 21% and 25.80%, respectively, from extreme storm-tide (1991-2007). In the most extreme storm-tide events considered in this study, the contribution of the nonlinear interaction resulted in smaller extreme water levels than those obtained through linear superposition of extreme storm-tide components. However, under certain conditions, the nonlinear interactions might result in higher storm-tides than those resulting from the linear superposition (e.g. storm of January 2000 at Cuxhaven and Sylt). The threat by extreme storm-tides might increase due to the change in future wind climate under emission scenario (A1B 2070-2100). The increase of the magnitude of the extreme surge-tide due to climate change is significant for the German Bight and the Netherlands coasts, while it decreases significantly for the middle of the North Sea and along the UK east coast. The temporal variations of the climate change signals for storm-tide at Cuxhaven and Sylt are assessed using the new hybrid model. The maximum ...


    Zugriff

    Download


    Exportieren, teilen und zitieren



    Titel :

    Combined Neural Network and CFD Modelling of Extreme Storm Surges ; Kombinier Modellierung Extreme Sturmfluten durch Neuronale Netzwerke und CFD



    Erscheinungsdatum :

    2015-09-08



    Medientyp :

    Hochschulschrift


    Format :

    Elektronische Ressource


    Sprache :

    Englisch



    Klassifikation :

    DDC:    6 / 62 / 627 / 690