Abstract:
Understanding and predicting coastal cliff retreat remains a challenging field in coastal geomorphology. Efforts to link process drivers to erosion are confounded by the complexity of processes operating, slow rates of morphological change, and the difficulty of direct field observations. Waves are generally considered to be a principle driver of cliff retreat, but few studies have conclusively shown how waves actually drive erosion. This thesis contributes to the field by improving understanding of how individual wave impacts interact with coastal cliffs. Wave impacts are classified based on the stage of transformation they have reached at the moment of impact; wave impact type falls on a continuum from broken to breaking to unbroken waves. Theoretical framing for the research is provided by an extensive history of engineering studies (laboratory experiments and numerical modelling) describing wave impacts on vertical coastal structures, as well as the relatively new field of cliff-top seismometry, which describes direct measurement of the amount of energy transferred from the wave into the cliff from the direct impact . The thesis describes a short-term (45 minute) set of field observations of wave impacts on a cliff in California, and a long-term (30 day) set of observations of impacts on a cliff at Onaero Bay in New Zealand. Analyses focus on wave impacts at the individual wave scale. Pivotal in facilitating this level of detail was the use of video recording to enable direct, real-time observations of the wave impact and resulting cliff response. By observing individual wave impacts with video and concomitant seismic response, the thesis ascertains the influence of wave impact type on peak ground motion magnitude. Wave impacts were initially divided into broken, breaking and unbroken waves at the moment wavecliff interaction takes place. Results highlight the dominance of wave impact class type in controlling cliff ground displacement. Interestingly, incident wave conditions are not the dominant control on cliff ground motion: the largest wave heights (Hs = 3.7 m) did not generate the highest peaks in ground motion. Instead, breaking wave impacts generally forced higher peak ground motion, regardless of incident wave height, when compared to broken or unbroken impact types. Breaking waves produced peak velocities in ground motion up to 7559 μm/s, compared to broken or unbroken waves, which produced maximum velocities of 1101 and 395 μm/s respectively. To further investigate the influence of wave impact class, 7447 individual impacts were subdivided into eight classes (WC1 – WC8) according to the stage in breaking transformation. The classification scheme was based upon an extensive body of engineering studies that have identified the primary controls on wave impact type on vertical coastal structures. The scheme incorporates an indication of the energy dissipated prior to impact, as well as the way the wave interacts with the cliff. The shape of the wave at impact is crucial, because it controls whether there is a pocket of air entrapped and compressed by the overturning wave. Analyses showed that breaking wave impacts (WC3-WC6) produced the highest ground motion compared to broken (WC1-WC2) and unbroken (WC7 and WC8) impacts. Breaking waves that entrapped a gas pocket at impact (WC4 and WC5) generated an average peak in ground displacement of 38 μm and 36 μm, whereas completely broken and unbroken waves produced peak displacement of around 2 μm on average. The variability within breaking, broken and unbroken impacts was attributed to the amount of aeration in the wave impact from bubbles or the size of the gas pocket. For the largest 1% of peak ground displacement values (>75 μm), 64% were WC4, 32% were WC5, 27% were WC3, and only 3% were unbroken or broken impacts (WC1-WC2, WC7- WC8) conducive to breaking impacts generating the most violent shaking events compared to broken and unbroken impacts. Spectral analysis was undertaken for each of the eight wave classes and distinctive spectral signatures for each class were revealed. This analysis was the first of its kind applied at the individual impact scale. The power spectral density for breaking wave classes is orders of magnitude higher than broken and unbroken impacts. WC4 impacts reached a peak of nearly 9 dB/Hz, compared to WC8 which reaches a maximum of 0.01 dB/Hz. Breaking waves (WC4 and WC5) also had the highest total integrated sum of energy transferred over a wave impact event: the median total energy transferred by WC4 and WC5 impacts was 117 dB/Hz and 121 dB/Hz, compared to 12 dB/Hz and 10 dB/Hz for WC1 and WC8 impacts. The highest energy peak for WC4 is nearly four orders of magnitude greater than for WC8. The thesis highlights the important and sensitive role of the ratio of wave height and water depth (H/h) in controlling wave impact class. There are time periods of dominant impact regimes where the H/h relationship is more conducive in producing a certain wave class or group of classes. The implications of this are that even relatively modest increases in future sea level are likely to fundamentally shift wave impact regimes at different cliff locations. In locations where sea level rise facilitates a greater number of breaking (WC4 and WC5) impacts, higher peaks in cliff ground motion are expected and this may increase rates of cliff recession in the future. In contrast, in some locations sea level rise may lead to a greater proportion of unbroken waves, decreasing erosion rates.