Investment planning models inform investment decisions and government policies. Current models do not capture the intermittent nature of renewable energy sources, restricting the applicability of the models for high penetrations of renewables. We provide a methodology to capture spatial variation in wind output in combination with transmission constraints. The representation of wind distributions with stochastic approaches or an extensive historic data set would exceed computational constraints for real world application. Hence we restrict the amount of input data, and use boot-strapping to illustrate the robustness of the results. For the UK power system we model wind deployment and the value of transmission capacity.