Improving Wave Climate Assessment with Wind Sea and Swell Components

For certain maritime operations there is an increasing demand for knowldedge about the co-existence of different wave systems. Design practices have shown that reliable directional data allow cost optimization in offshore facilities by reducing the inconsistencies in design criteria due to omni-directional approaches. The main goal for this dissertation is to improve the traditional assessment of wave climates by introducing wind sea and swell systems as regular components. Three independent datasets for a location on the Belgian Continental Shelf has been extensively examined during the period 1999-2008. The first dataset consists of buoy measurement records. The other two datasests are two wave model hindcasts, the ERA-INTERIM re-analysis from ECMWF on the one hand and in house simulations using the SWAN wave model on the other hand. Careful pre-processing was carried out to have a homogeneous dataset. The development of a wave partitioning scheme is considered a main achievement.The MuSeas toolbox has been built in Matlab environment and has proven to be a reliable and flexible tool for analyzing multiple seas. In particular, effort has been made to obtain an analysis routine capable of dealing with different type of input and formats and with a high degree of automatization. An inter-comparison between datasets and wave components for the mean climate was carried out to have a measure of the interchangeability between observations and model results. The hindcasts showed satisfactory representations of the total and of the wind sea component, but it was not so adequate for the swell component. Multivariate analysis for the partitioned spectra was carried out for the mean wave climate. A novel approach visualizes multiple wave systems by way of mosaic plots derived from contingency tables. These plots display characteristic patterns of wind sea and swell systems for a specific location. Another practical application deals with combinations of multiple seas that are considered hazardous for maritime navigation as in the case of crossing seas. Also the extreme behavior of wind sea and swell wave components was investigated both from a stationary and non-stationary point of view. Stationary modeling assumes that the distribution parameters remain constant over time, whereas the non-stationary model involves also the seasonal fluctuations. And finally the effects due to gaps in the data time series and due to directionality are investigated.status: publishe