Wind dataset assessment and energy estimation for potential future offshore wind farm development areas on the Scotian Shelf

Abstract. The Scotian Shelf is one of the top wind regimes in the world. In order to assess the wind energy of the potential wind farms over the shelf, in this study, we first assessed the uncertainties of four commonly used wind datasets: ERA5, CFSv2, NARR, and HRDPS, by comparing them against observational wind data distributed at both nearshore and offshore sites. The assessment indicates that the root-mean-square error of the datasets varies between 1.6 m/s and 2.4 m/s in wind speed and between 24.6° and 36.4° in wind direction. HRDPS performs better at the near-shore sites, while ERA5 is more accurate at the offshore sites. We then estimated the wind energy potential of six wind farms on the shelf using ERA5 and HRDPS. The estimation shows that wind energy varies seasonally, the energy in summer 55 % lower than that in winter. The uncertainties in wind datasets enhance the variation of the wind energy production, up to 28 % in winter and 55 % in summer. The energy output is sensitive to turbine spacing due to wind wakes, which reduce energy by 17 % to 26 % in winter and by 40 % to 55 % in summer, depending on the relationships between wind speeds, wind directions, and the specific layout of the wind farms. This strong variation in wind energy output suggests that a more feasible operational method should be used to balance energy production and usage.