Long-term wind resource assessment for small and medium-scale turbines using operational forecast data and measure-correlate-predict

Output from a state-of-the-art, 4 km resolution, operational forecast model (UK4) was investigated as a source of long-term historical reference data for wind resource assessment. The data were used to implement measure-correlate-predict (MCP) approaches at 37 sites throughout the United Kingdom (UK). The monthly and hourly linear correlation between the UK4-predicted and observed wind speeds indicates that UK4 is capable of representing the wind climate better than the nearby meteorological stations considered. Linear MCP algorithms were implemented at the same sites using reference data from UK4 and nearby meteorological stations to predict the long-term (10-year) wind resource. To obtain robust error statistics, MCP algorithms were applied using onsite measurement periods of 1-12 months initiated at 120 different starting months throughout an 11 year data record. Using linear regression MCP over 12 months, the average percentage errors in the long-term predicted mean wind speed and power density were 3.0% and 7.6% respectively, using UK4, and 2.8% and 7.9% respectively, using nearby meteorological stations. The results indicate that UK4 is highly competitive with nearby meteorological observations as an MCP reference data source. UK4 was also shown to systematically improve MCP predictions at coastal sites due to better representation of local diurnal effects.