The Role of Input Assumptions and Model Structures in Projections of Variable Renewable Energy: A Multi-Model Perspective of the U.S. Electricity System

Recent wind and solar deployment in the United States has grown at unprecedented rates, but the extent to which this growth will continue is a key question. Energy-economic planning models are often used to develop future projections of these technologies, and although it is well-understood that model outcomes are sensitive to assumptions about technology developments and market conditions, it is less clear how underlying model structures might impact outcomes. For wind and solar technologies, model structures might be an even larger driver due to the complexity required to represent their variability, uncertainty, and location-dependence. Here, we apply three state-of-the-art electric sector planning models to evaluate variable renewable generation under different conditions with both harmonized and native input assumptions to isolate the impact that model structures might have on deployment outcomes. We find growth in wind and solar deployment in all models under a business-as-usual scenario (without new policies), where the median value for variable renewable energy generation share is found to be 27% by 2050—but with a wide range of estimates (18–43%), types of renewable technologies deployed, and regional distributions across models. With harmonized assumptions about technology costs and market conditions the range narrows slightly, but a large range in penetration outcomes remains (24–41% by 2050) suggesting that model representations play significant roles in projections of future wind and solar deployment. We also assess the implications of achieving 55% wind and solar penetration by 2050—to assess potential challenges at higher penetration levels—and the magnitude of variable generation under a carbon-constrained future. Finally, we identify research needs to better assess the future deployment of variable renewable energy.