Hide/Show Apps

Assessment of performance of interpolation methods for temporal resolution of wind data on wave modelling

Ak, Gülçe Hazal
In this thesis, the performance of different interpolation techniques for wind hindcasting have been assessed and a computer model based study have been performed on its effects on wave hindcasting. Climate Forecast System Reanalysis (CFSR) dataset between 1979 and 2010 is used for long term time series analysis and extreme event analysis in Black Sea region. The aims are to obtain new dataset using the existing dataset, to determine the reliability of statistical analysis of winds and to evaluate the error margins in numeric wave models. W61, a parametric wave hindcasting model has been used for long term and extreme wave analysis and WAVEWATCH III has been used for further analysis of 3 extreme events within the Black Sea. Several statistical parameters are used to evaluate and compare the interpolation methods including Mean Absolute Error, Coefficient of Variation Root Mean Square Error and Normalized Bias. The wave and statistical analysis results indicated that the performances of the interpolation methods for wave analysis are not compatible with the wind analysis results. Also, performances of the interpolation methods varies through the analyzed 25 points within the Black Sea for both long term and extreme wave analysis.