Posters

Abstract

Measure-Correlate-Predict (MCP) is a well-known method used to adjust short-term wind speed measurements to predict the long-term wind resource at a potential new wind farm site (target site). It establishes a relationship between the wind speed data measured at a target site and suitable data from an appropriate reference source where concurrent wind speed and direction data are available as well as long-term data. This relationship (for example, a linear regression by direction) is then used to scale the short-term measured data to the period for which the long-term data are available assuming a stable climatology. In the past, data from established meteorological stations were used to provide such a reference. However, using data from such sites for this purpose has drawbacks such as poor data quality, changes to instrumentation and changes in site exposure, e.g. due to the growth or removal of trees. Furthermore, a suitable close-by reference meteorological station where the measured wind speed data are well correlated to those of the target site may not be available. More recently, reanalysis products have become available where numerical weather prediction models are used to provide a hindcast estimate of numerous meteorological variables on a regular 3D grid at up to an hourly temporal resolution based on the assimilation of a comprehensive database of past observations. These products are well validated and do not suffer from many of the drawbacks of using data from meteorological stations. However, little research has been done regarding how the use of such modelled data in MCP affects the achievable wind resource prediction accuracy. This research set out to investigate the accuracy of the MCP procedure comparing the use of wind speed data from meteorological stations with equivalent data from the ERA5 reanalysis as a long-term reference source. The accuracy of the MCP technique using these two types of reference datasets is assessed through the ability to predict wind speed values measured at 35 sites located in different terrain types. This study found that ERA5 reanalysis data can serve as a reliable alternative to observed meteorological station data. Generally, using ERA5 reanalysis data as a reference source always led to more accurate predictions then a meteorological station reference source if the Pearson correlation coefficient between the target and meteorological station wind speed values is lower than 0.8, and for all the offshore target sites. If the Pearson correlation coefficient between the target and meteorological station wind speed is higher than 0.9, the achieved accuracy using either the meteorological station wind speed or ERA5 wind speed data as a long-term reference is similar and depends on specific site conditions.