Historical wind deployment and implications for energy system models

A critical parameter in modeling studies of future decarbonized energy systems is the potential future capacity for onshore wind power. Wind power potential in energy system models is subject to assumptions regarding: (i) constraints on land availability for wind deployment; (ii) how densely wind turbines may be placed over larger areas, and (iii) allocation of capacity with respect to wind speed. By analyzing comprehensive databases of wind turbine locations and other GIS data in eleven countries and seventeen states in Australia, Canada, and the US; all with high penetration levels of wind power, we find that: i) large wind turbines are installed on most land types, even protected areas and land areas with high population density; ii) it is not uncommon with a deployment density up to 0.5 MW/km2 on municipality or county level, with rare outlier municipalities reaching up to 1.5 MW/km2 installed capacity; and iii) wind power has historically been allocated to relatively windy sites with average wind speed above 6 m/s. In many cases, allocation methods used in energy system models do not consistently reflect actual installations. For instance, we find no evidence of concentration of installations at the windiest sites, as is frequently assumed in energy system models. We conclude that assumptions made in models regarding wind power potentials are poorly reflective of historical installation patterns, and we provide new data to enable assumptions that have a more robust empirical foundation.