Detailed and reliable spatiotemporal characterizations of turbine hub height wind fields over coastal and offshore regions are becoming imperative for the global wind energy industry. Contemporary wind resource assessment frameworks incorporate diverse multiscale prognostic models (commonly known as mesoscale models) to dynamically downscale global-scale atmospheric fields to regional-scale (i.e., spatial and temporal resolutions of a few kilometers and a few minutes, respectively). These high-resolution model solutions aim at depicting the expected wind behavior (e.g., wind shear, wind veering and topographically induced flow accelerations) at a particular location. Coastal and offshore regions considered viable for wind power production are also known to possess complex atmospheric flow phenomena (including, but not limited to, coastal low-level jets (LLJs), internal boundary layers and land breeze–sea breeze circulations). Unfortunately, the capabilities of the new-generation mesoscale models in realistically capturing these diverse flow phenomena are not well documented in the literature. To partially fill this knowledge gap, in this paper, we have evaluated the performance of the Weather Research and Forecasting model, a state-of-the-art mesoscale model, in simulating a series of coastal LLJs. Using observational data sources we explore the importance of coastal LLJs for offshore wind resource estimation along with the capacity to which they can be numerically simulated. We observe model solutions to demonstrate strong sensitivities with respect to planetary boundary layer parameterization and initialization conditions. These sensitivities are found to be responsible for variability in AEP estimates by a factor of two. Copyright © 2013 John Wiley & Sons, Ltd.