Ocean waves driven by wind stress—wind-waves—are familiar and impressive, bringing benefits such as renewable energy and surfing, but frequently act as a hazard to coastal and offshore industries. For example, recent studies have identified extreme levels of erosion on European and North American coasts, driven by increasing storm intensty in the midlatitudes (Barnard et al. 2017, Masselink et al. 2016, Harley et al. 2017). However, tropical cyclones also drive extreme wave climate, and growing evidence for the connection between TC activity and increasing ocean temperature, motivate investigation of possible future changes.
There is huge research interest in observational records of precipitation that spans environmental and hydrological modeling, climate change and resource and risk management (Groisman et al. 1999, Krishnamurthy & Shukla 2000, Alexander et al. 2006, Schär et al. 2016). In order to make data access more practical, and also to incorporate post-processing of the raw observations, a large number of geographically gridded precipitation products have been developed. However, it remains unclear how well extremes are represented, or whether there is agreement in their representation. Analysis using bivariate extreme value theory, based on a copula approach, identifies temporal discrepancies.
Research in the geophysical sciences employs a plethora of numerical models, many of which are extremely expensive, particularly when run at resolutions high enough to resolve complex phenomena. As a result, sensitivity and uncertainty analysis is challenging because models cannot be run in numbers required to perform robust analysis, particularly for extremes. In order to, at least partially, circumvent these challenges, methods of applied statistics can be used obtain fast "surrogate" models