As the demand for global renewable energy grows, so does the demand for more efficient energy conversion machines and better wind resource assessment. The need to convert as much energy as possible with little cost remains the biggest challenge. In the wind energy sector, the quantity of the resource “wind” is not hard to locate, as with current ground and space based remote sensing technologies, and climate reanalysis techniques, the mapping of average wind speeds across the globe is achievable. The difficulty lies in identifying the “quality” of the wind resource. “Quality” is the measure of the time variant properties of the wind, and time scale here does not represent seasonal, monthly, or the daily variability, but rather the changes within hours, minutes, seconds, and sub‐second periods. Wind possesses a highly unpredictable, and non‐universal character, which is referred to as turbulence. These intermittencies in the wind speed create variable mechanical loads on the structure of wind turbines leading to fatigue, and ultimately failure. Identifying site specific qualities of the wind resource is very crucial in the design and selection process of the wind turbine. Physical theories explaining wind turbulence phenomena over flat terrain have been critiqued and tested by observations, and in general, have achieved reasonable success in explaining surface layer wind dynamics that can be applied universally. This universality, and the extrapolation of flat terrain theories to complex terrain applications, breaks down most of the time due to the newly recognized spatial and temporal spectrum of interaction modes, mechanically and thermodynamically, with the surrounding complex terrain. In terrain as found in New Zealand, most of the wind farm development is carried out over complex terrain, with ridge top and mountainous installations. In this study, an experimental campaign was carried out over a coastal ridge top, proposed for wind farming, to investigate mean and turbulent wind flow features significant for wind turbine selection and placement across the ridge. The steep sloped faces of the ridge, high wind speeds and its proximity to the sea made this location ideal for a benchmark investigation site. Ultra‐sonic ii anemometers, a sodar (sound detection and ranging) wind profiler, and high resolution LES (large eddy simulation) numerical modelling were all utilized separately and in an interconnected way to provide a comprehensive analysis of the wind dynamics over the ridge top. The three principal components of the investigation were: the effect of the upstream topography and the thermal circulation associated with the proximity to the sea on the observed and modelled wind shear vertical profile; the role that the near upwind terrain plays in shaping the turbulence energy spectrum and influencing the predicted spectrum, ultimately affecting isotropy in the flow field and turbulence length scales; turbulence advection from far topography, and the role that far upwind terrain plays in altering the wind turbulence in a measurement area or at a single point. Results showed that the thermal wind circulations and upstream steep topography could dictate the wind shear profile, and consequently have a large impact on wind turbine height selection and placement. The sodar proved to be a very useful tool in identifying vertical shear zones associated with effects of steep upstream terrain, vertical mixing of horizontal momentum, and thermal circulation from the local sea breeze. In complex terrain, the added multi‐directional perturbations from the underlying roughness redistribute the statistical variations (measured by variances) in the three spatial dimensions. Isotropy, based on measured variances, was attained for both sites on the ridge. Isotropy also held true for the energy spectrum via Fourier analysis of the high temporal resolution data, but not for both sites. In general, local isotropy can be attained in cases of higher wind speeds and increased terrain relief. Measured spectral ratios did not converge to the limit suggested by the local isotropy hypothesis. These results identify contradictions in assessing the turbulence isotropy in both real space (statistically through variances) and Fourier space (through power spectrum analysis), which suggests caution in deriving or interpreting turbulence information for wind turbine design and selection. iii 2D‐LES experiments showed that turbulent kinetic energy (TKE) can attain long range memory of underlying terrain, which can then react accordingly with upcoming terrain. Under the high wind speed scenarios, which are suitable for wind farming, and over relatively complex terrain, the flow retained some aspects of terrain information at least 30H (H is the terrain height) upstream and downstream of the terrain. In general, as the turbulence field travels over new terrain it tends to increase in intensity downstream of that feature. The newly modified TKE field acquires geometric features from the underlying terrain; mainly these features register as amplifications in the wave structure of the field at wavelengths comparable to the height of the underlying terrain. The 2D‐LES sensitivity experiments identified key areas of high mean wind speed and turbulence in relation to terrain effects, all of which should be taken into consideration when thinking of locating a wind farm in such areas.