The desire to reduce dependence on fossil fuels is resulting in numerous policy incentives for increased renewable energy sources within the power grid. Because wind generation is arguably the most affordable per MWh of the renewable energy sources it is growing nearly as quickly as conventional generation techniques. Due to this significant increase in wind penetration levels, numerous largescale wind integration studies have been produced to determine the reliability impacts of large-scale wind power. Using data from two large US wind interconnection studies, this thesis provides evidence that mesoscale meteorological models under-predict the variability in wind data particularly on short time scales, indicating that data from mesoscale meteorological models need to be used with caution for some types of analyses. These types of analyses include most notably regulating reserves, which are used to rebalance supply and demand on a second-by-second bias. This thesis will also describe and evaluate a new method for jointly quantifying the amount of spinning and regulating reserves required to meet reliability requirements within a balancing area with significant amounts of wind power using high resolution wind data. The method is based on jointly minimizing dispatch costs and reserve allocations, across two time scales (seconds to minutes, and minutes to hours) to satisfy North American Electric Reliability Corporation (NERC) Area Control Error (ACE) requirements.