On weather and waves : applications to coastal engineering.

Abstract

Shoreline erosion in response to extreme wave events can be severe. The reduction in beach width leaves development within the hinterland exposed and vulnerable to future wave attack. Wave climates are a fundamental driver of coastal erosion and changes to wave height, direction and period can severely impact a coastline. These changes are directly linked to changes within the principle drivers of wave climates namely synoptic scale atmospheric circulation. The links are complex and if they can be clarified they can be used to provide insight into wave climates and improve the evaluation of future climate scenarios. The coupling between atmospheric circulation and wave climates provides a tool for risk assessment that is strongly based on fundamental physical processes. This study is focused on exploring this relationship and its effect on coastal vulnerability. A statistical classification algorithm is utilized to explore the relationship between synoptic scale circulation patterns and regional wave climates. The algorithm is fully automated and discrete atmospheric patterns are derived through an optimization procedure. It is driven to an optimal solution through statistical links between regional wave climates and atmospheric circulation patterns (CPs). The classification is based on the concept of fuzzy sets and differs from standard classification techniques. It employs a "bottom–up" approach as the classes (or CPs) are derived through a procedure that is guided by the wave climate. In contrast existing classification techniques first explore the atmospheric pressure space while links to the variable of interest are only made post classification. The east coast of South Africa was used as a case study. Wave data off the Durban coastline were utilized to evaluate the drivers of the wave climate. A few dominant patterns are shown to drive extreme wave events. Their persistence and strong high– low coupling drive winds toward the coastline and result in extreme wave events. The sensitivity of the algorithm to key input parameters such as the number of CP classes and temporal resolution of the data was evaluated. The Shannon entropy is introduced to measure the performance of the algorithm. This method benefits from incorporating the link between atmospheric CPs and the wave climate. A new stochastic wave simulation technique was developed that is fundamentally based on the CPs. This technique improves the realism of stochastic models while retaining their simplicity and parsimony relative to process-based models. The simplicity of the technique provides the framework to evaluate coastal vulnerability at site specific locations. Furthermore the technique was extended to evaluate changes in wave behaviour due to climate change effects.