This study presents the south-west Black Sea wave climatology based on a downscaling approach of a long-term 31-year SWAN model wave hindcast using telescoping nested grids. At all domains the SWAN model is forced with the CFSR winds. Sensitivity tests are conducted on domain size, computational resolutions, the physical formulations and their adjustable coefficients for deep water source terms, time step of non-stationary calculation, and wind forcing for all domains. For each nested grid the physical and numerical settings were determined separately by calibration against wave buoy measurements at six locations (Gelendzhik, Hopa, Sinop, Gloria, Filyos, and Karaburun) in appropriate domains. Model validation is also conducted for the long-term data using the unused measurements in the calibration. Using the calibrated nested models, a 31-year long-term wave hindcast is conducted. Two-hourly sea state parameters of significant wave height (Hm0), wave energy period (Tm-10), spectral period (Tm01), zero-crossing period (Tm02), peak period (Tp), wind speed components, and mean wave direction (DIR) were collected over three sub-grid domains. Using this database normal and extreme wave conditions in the three sub-grid domains were determined. Finally, extreme waves with different return periods were determined and compared with those presented in the Wind and Deep Water Wave Atlas of Özhan and Abdalla (2002). The present study demonstrates the sensitivity of the SWAN model towards different GEN3 physics options and its adjustable whitecapping parameter Cds and time step of the non-stationary calculations. It is shown that the developed wave model set-up with a nested grid system performs quite satisfactorily and storms are also well-captured. This study yields higher extreme waves in the western part of our area of interest and lower extremes in the eastern part in comparison with those of the presently used Wind and Deep Water Wave Atlas.@en