Diffraction of regular waves, focused wave groups, and random seas by arrays of vertical bottom mounted circular cylinders is investigated using theoretical, computational, and experimental methods. Free surface elevation η is the defining variable used to test the potential theory developed. In addition, the nonlinearity of focused wave groups is investigated through the Creamer nonlinear transform and analysis of numerical wave tank data. Linear focused wave group theory is reviewed as a method for predicting the probable shape of extreme events from random wave spectra. The Creamer nonlinear transform, a realistic model for steep waves on deep water, is applied in integral form to simulate nonlinear focused wave groups. In addition, the transform is used to facilitate analysis of nonlinear wave-wave interactions within focused wave groups from a uni-directional numerical wave tank developed at Imperial College London. Experiments in an offshore wave basin at HR Wallingford are designed to measure free surface elevation at multiple locations in the vicinity of a multicolumn structure subjected to regular and irregular waves for a range of frequencies and steepness. Results from regular wave data analysis for first order amplitudes are compared to analytical linear diffraction theory, which is shown to be accurate for predicting incident waves of low steepness. However, second and third order responses are also computed, and the effects in the vicinity of a second order near trapping frequency are compared to semi-analytical second order diffraction theory. Analytical linear diffraction theory is extended for application to focused wave groups and random seas. Experimental irregular wave data are analysed for comparison with this theory. Linear diffraction theory for random seas is shown to give an excellent prediction of incident wave spectral diffraction, while linear diffraction theory for focused wave groups works well for linearised extreme events.