Extreme value analysis: wave data

Abstract

Estimates of the m-year return value of significant wave height ─ the value which is exceeded on average once every m years ─ are needed for the safety control and design of ship, offshore, and coastal structures, and for the mapping of flood risk areas. The WMO Guide to Wave Analysis and Forecasting aims at providing guidance on how to obtain those estimates. In the design of ships and offshore platforms 1/20-yr to 1/100-yr return values are often used. In the control of the safety of the Netherlands sea defenses return values of up to 1/10,000-yr are used. In the mapping of food risk area in the United Kingdom 1/1,000-yr return values are used. The longer time series of significant wave height available come from hindcasts and usually cover no more than 50 years, meaning that one generally needs to extrapolate well beyond the range of the available data and thus resort to extreme value analysis to obtain the required return value estimates. In this report we begin by describing and discussing approaches that can be used to estimate such return values in Chapter 2. Approaches based on extreme value theory as well as ad hoc methods are considered. We then present in Chapter 3 some worked examples using two time series of significant wave height measurements, one in deep and the other in shallow waters. In Chapter 4 we provide an inventory of software packages available to carry out extreme value analyses. We finish in Chapter 5 with some guidelines / recommendations.