Renewable energy grid connection is hampered by transmission capacity limitations and public opposition to new transmission development. This paper presents a methodology to find the optimal positions on an existing transmission system network to connect 'firm' wind capacity to reach desired renewable energy penetration targets in a secure, least-cost manner. The methodology accounts for geographical statistical dependencies of individual bus load and wind farm power outputs, as well as the temporal dependencies of the conventional plant unit-commitment process on total system load and wind patterns. This is accomplished using a probabilistic load flow technique based on DC load-flow and recorded load and wind time series. A discretised model of the resultant multi-variate probability density function is used to define line flow constraints in a linear programming optimization model. The algorithm objectively allocates wind capacity with respect to the wind resource and transmission capacity in each area.