Wind en golven: onderzoek van een optimaliserings-benadering voor het schatten van parameters in de brontermen = Wind and waves: investigation of an optimization approach to parameter estimation
Monbaliu, J. (1992). Wind en golven: onderzoek van een optimaliserings-benadering voor het schatten van parameters in de brontermen = Wind and waves: investigation of an optimization approach to parameter estimation. PhD Thesis. Katholieke Universiteit Leuven (KUL): Leuven. XXXVII, 196 pp.
The atmosphere sees the ocean, with sometimes huge waves, as a much smoother surface, compared to its passage over land where it is slowed down much more, even if obstacles are small compared to the size of ocean waves. A good observer notices right away that waves travel with the wind and consequently do not give much resistance. The details of how much resistance there actually is and its variation with wind speed and parameters defining the wave conditions has been a hot topic for the last 35 years. The correctness of the roughness formulation is a crucial question in the understanding of the air-sea interface. It has major implications in wave modelling, in storm surge modelling and also in modelling of the atmospheric boundary layer above the ocean. Researchers now agree that the roughness is not only a function of the wind speed, but also of the age of the waves. Young wind seas give more drag resistance than old wind seas. Mainly waves at frequencies beyond the peak frequency contribute to the roughness. We propose to use the Donelan constant, representative for the equilibrium range of a wind sea spectrum, as a direct measure for the nondimensionalized roughness of the sea surface. To obtain a better view and understanding on the working of the source terms in a third generation wind wave model, a framework has been build for tuning unknown parameters in the formulation of the forces involved in wave generation and wave dissipation through whitecapping. In the framework a standard optimization routine from the widely available NAG-Fortran subroutine library is used to minimize a cost function. The cost function is defined as a weighted sum of squares to take advantage of the special structure of this formulation. Measured physical characteristics such as wave energy, and peak frequency of the waves are expressed in function of fetch through the JONSWAP or the Kahma and Calkoen growth curve laws. In the optimization exercise, we try to reproduce the fetch evolution of the total energy by tuning two elected parameters in the wind input and/or dissipation source terms. Such a framework is useful in increasing our understanding of individual source terms.
All data in the Integrated Marine Information System (IMIS) is subject to the VLIZ privacy policy
