In wind turbine design, external conditions to be considered depend on the intended site for the planned installation. Wind turbine classes, defined in terms of wind speed and turbulence parameters, cover most sites and applications. In the International Electrotechnical Commission's (IEC's) 61400-1 standard, there is a design load case that requires consideration for ultimate loading resulting from extreme turbulence conditions. Since site-specific wind conditions should not compromise the structural integrity of turbine installations, at some sites where class-based design may not apply, there is sometimes a need to establish extreme turbulence (50-year) levels as part of site assessment by making use of measurements. This should be done in a manner consistent with class-based design where the extreme turbulence model (ETM) provides 50-year turbulence standard deviation (σ) value as a function of the ten minute average hub-height wind speed, V. For one site in Germany and three contrasting terrain sites in Japan, wind velocity data are used to establish 50-year ETM levels. The inverse first-order reliability method (IFORM) is applied with 10 min data for this purpose. Sometimes, as in assessing wind farm wake effects, analysis of turbulence levels by direction sector is important because normal and extreme turbulence levels can vary by sector. We compare ETM levels by sector for the Hamburg, Germany site. The influence of terrain complexity on ETM levels is also of interest; the three sites in Japan have contrasting terrain characteristics—referred to as flat, hilly, and mountainous. ETM levels are compared for these three terrain types. An important overall finding of this study is that site-specific ETM levels can greatly exceed levels specified in the standard for class-based design.