Using map algebra to explain and project spatial patterns of wind energy development in Iowa

Rapid ongoing development of wind power raises the question of where new wind turbines will be placed. This study uses locational decisions made through 2010 to develop a logistic regression model of wind turbine location among one square kilometer cells in Iowa, the U.S. state with the highest density of wind turbines. An 8-variable model correctly predicts 85 percent of cells. Wind energy density at 50 m and 80 m height are positively related to wind power; so also is population density within a 200 km radius and cropland. Distance from mid-voltage power lines and interstate highways are negatively related as is an airport within 5 km and population density within a 50 km radius. Using map algebra, the logit model generates a map of the likelihood of wind energy development. Locations that would most benefit from augmented electrical transmission are also identified. This form of empirical locational analysis can thus help predict and guide wind power development where spatial data to calibrate a high-fit logit model are sufficient.

► Raster-based logistic regression can correctly predict 85 percent of 1 km raster cells containing wind turbines in Iowa.
► This method can also project a most likely spatial distribution of further wind power development.
► This method can also identify where transmission line development would most improve wind turbine suitability.
► Wind turbines are associated with high wind speeds, nearness to power lines, interstate highways, and cropland.