1 Spatial Dependence Modeling of Wind Resource under 2 Uncertainty Using C - Vine Copulas and Its Impact on 3 Solar - Wind Energy Co - Generation 4

1 University of Waterloo, Ontario, Canada N2L 3G1; {a22naray, ponnu}@uwaterloo.ca 6 2 University of Massachusetts Lowell, Massachusetts, USA 01854; [email protected] 7 * Correspondence: [email protected], Tel.: +1-519-888-4567 x31450 8 Abstract: Investments in wind and solar power are driven by the aim to maximize the utilization of 9 renewable energy (RE). This results in an increased concentration of wind farms at locations with higher 10 average wind speeds and of solar panel installations at sites with higher average solar insolation. This is 11 unfavourable for energy suppliers and for the overall economy when large power output fluctuations occur. 12 Thus, when evaluating investment options for spatially distributed RE systems, it is necessary to model 13 resource fluctuations and power output correlations between locations. In this paper, we propose a 14 methodology for analyzing the spatial dependence, accurate modeling, and forecasting of wind power 15 systems with special consideration to spatial dispersion of installation sites. We combine vine-copulas with 16 the Kumaraswamy distribution to improve accuracy in forecasting wind power from spatially dispersed 17 wind turbines and to model solar power generated at each location. We then integrate these methods to 18 formulate an optimization model for allocating wind turbines and solar panels spatially, with an end goal 19 of maximizing overall power generation while minimizing the variability in power output. A case study of 20 wind and solar power systems in Central Ontario, Canada is also presented. 21