Presentation Description: The wind resource assessment and energy output predictions currently rely on detailed numerical simulations that might not be accurate enough when realized over flat and/or complex topography. Two of the main challenges encountered today in microscale wind simulations are, firstly, the fact that most surface treatments are based on similarity theories developed for ideal flat terrain (i.e. wall functions) and, secondly, the computational cost of high resolution turbulent flow modeling over complex sites. As a contribution to obtain a wind flow solver with higher accuracy and computational efficiency, an atmospheric turbulence model is proposed by unifying the k-omega SST unsteady Reynold averaged Navier-Stokes (URANS), based on Menter et al. (2003), and the Simplified Improved Delayed Detached-Eddy Simulation method (SIDDES), introduced by Patel and Yoon (1996). Based on this hybrid model, wind flow can be resolved down to the wall, if a proper mesh is used without significant overhead, and the turbulence model equations directly account for surface roughness which closely replicates real flow phenomena. An extensive analysis of canonical and real flat terrain flows is carried out to validate and outline the capabilities of this novel modeling approach. Subsequently, simulations of the neutrally stratified atmospheric boundary layer flow over heterogeneous terrain for a real field test are presented. The k-omega SST-SIDDES hybrid model has proven to be a suitable approach for predicting realistic wind behaviour without relying on flat terrain assumptions, thus, overcoming some of the current wind modeling limitations that will allow improved and reliable wind resource assessments for both on and off-shore projects.