A series of experiments were performed to determine the velocity field and other characteristics of large amplitude shallow water waves. The experimental results were compared with the predictions of a variety of wave theories including those commonly used in engineering practice. While no theory was found exceptionally accurate, the cnoidal wave theory of Keulegan and Patterson appears most adequate for the range of wavelengths and water depths studied.

A comparison of the conditions of similitude for movable bed scale models of rivers and estuaries on one hand, and beaches and shorelines on the other hand, is presented It is demonstrated that despite the fact that the knowledge in sediment transport by wave action is less advanced than in the case of steady current, the conditions of similitude for beaches are less stringent than for rivers In particular, the effect of relative roughness is comparatively unimportant in the case of beaches, while the necessity of similitude of head loss imposes an imperative condition in the case of scale models of rivers An introduction to a natural law of distortion for beaches is presented in analogy with the Lacey condition for rivers.

A mathematical model for long term shoreline evolution is developed. The combined effects of variations of sea level, wave refraction, wave diffraction, loss of sand by density currents during storms, by rip currents and by wind, bluff erosion and berm accretion as well as effects of man-made structures such as long groin or navigational structures and beach nourishment are all taken into account. A computer program is devloped with various subroutines which permit modification as the state-of-the-art progresses. The program is applied to a test case at Holland Harbor, Michigan.

The effect of sample size on confidence band in predicting the extreme wave height is related to the return period and the duration of the sample record length. An estimate of errors resulting from various methods of data acquisition is given. The uncertainties corresponding to various number of years of observation and various measurement errors are analyzed for various return intervals. It is shown that data accuracy and record length are equally important in long term wave predictions. At the present time, the determination of extreme event benefits more from relatively less accurate long-term hindcast calculation rather than short-term high quality measurements. In the long run, a long-term accurate measurement program is imperative if more definite descriptions of extreme events are sought.