WP 2

Reliable predictions of storm-tides and nearshore waves/currents are of vital importance for beach nourishment projects., but the mutual nonlinear interactions of all constituents of storm-tides and their contributions to extreme water levels cannot yet be fully solved by conventional hydrodynamic models alone (Oumeraci, 2004). These nonlinear interactions generally results in lower water levels than those obtained through linear superposition of extreme storm-tide components (Goennert and Gerkensmeier, 2015), but might also result in higher storm-tides. Artificial Neural Networks (ANNs) have been successfully applied to forecast storm-tides and waves (Bajo & Umgiesser, 2010), but they are limited to short-term predictions. Recently, Tayel (2015) developed a new approach by combining state-of-the art hydrodynamic models and dynamic neural networks (NARX). The following capabilities of this approach were exemplarily illustrated for two sites in the North Sea: (i) Reproduction of observed storm-tides over 8 years, including the contribution of the nonlinear interactions of the storm-tide constituents, (ii) Filling (water level) data gaps at one site by using data from a neighboring site, and (iii) Prediction of storm-tide development under present and future climate conditions.

Arbeitspaket 2 wird am Leichtweiß-Institut für Wasserbau der Technischen Universität Braunschweig durchgeführt.