The Influence of Wind Field Data Simplification on Wind-Wave Hindcasting in Great Lakes

Due to the sparseness of the available data for wind-wave hindcasting, temporal and spatial gaps exist in wind input data. Also, for the sake of convenience and reduction of computational efforts, the available wind field data may be simplified when used in numerical simulations of wave generation. Although this simplification is computationally economical, the wave characteristics obtained from the numerical simulations may be compromised. This study investigates the extent of influence for wind data simplification on the characteristics of wind-generated waves by considering two case studies. SWAN (SWAN Group, [25]), a third generation wave prediction model specialized for coastal and inland waters, are applied to hindcast wave regime. Lake Oroumieh in Iran and Lake Michigan in USA are considered as case studies. The effect of temporal wind variations are assessed by comparing predicted wave heights using the wind-rose and wind time-series as input data for Lake Oroumieh. A noticeable difference in the maximum significant wave height and period is observed between the results of the two prediction cases. The simulation results of the wind-rose case predict a wave height approximately 100% higher than the case with wind time-series data. In Lake Michigan the influence of the spatial variation of wind input data has been evaluated by comparing the predicted wave heights from simulation of surface spatially varied wind field to that of the constant wind over the whole lake area. Therefore in practical designs, considering local wind effects and storms by using satellite data catchments, buoys, ships reports and meteorological stations can play a significant role in reducing the final project costs.