In the wind resources assessment methods of regional wind farms

the observation data of reference stations is not sufficiently utilized by traditional measure-correlate-predict (MCP) methods to establish the mapping relationship between the wind power density of reference stations and the target station. Therefore

the forecasting accuracy of the long-term wind energy resources at the target station is not high. Based on the traditional MCP combination theory

two different MCP models are established with support vector regression (SVR). In addition

the traditional MCP models considering the characteristics of wind speed and direction are taken as the benchmarks to verify the effectiveness and accuracy of the proposed models. Results show that the determination coefficient of the proposed MCP method for wind power density prediction of the target station is higher than 0.9

and the forecasting accuracy and feasibility are significantly improved in comparison with those of traditional MCP models. Therefore

the proposed model can be applied to assess the wind energy resources of regional wind farms.