The accurate prediction of wake dynamics behind spire-generated atmospheric boundary layers is a critical challenge in computational fluid dynamics (CFD), particularly when resolving vertical and lateral velocity components under varying mesh resolutions. This study investigates the sensitivity of numerical wake characteristics to computational grid refinement in a spire-induced turbulent flow environment. Spires are widely employed in wind tunnel simulations to reproduce realistic atmospheric boundary layer (ABL) conditions, yet their interaction with downstream wake structures remains highly dependent on discretization strategies.

A structured CFD framework is developed to evaluate the influence of mesh density on velocity field accuracy, focusing on wake development, turbulence intensity distribution, and flow anisotropy. The study integrates theoretical ABL modeling principles established in classical wind engineering research (Armitt & Counihan, 1968; Counihan, 1969), while incorporating modern spire-based flow generation methodologies that ensure realistic boundary layer reproduction. Additionally, insights from experimental and numerical investigations on spire-induced flow fields are considered to validate simulation reliability (Fitriady, Rahmat, & Mohammad, n.d.).

The results demonstrate that finer mesh resolutions significantly enhance the fidelity of vertical velocity gradients and lateral wake diffusion, while coarse meshes tend to underestimate turbulence anisotropy and wake recovery length. The findings provide critical implications for CFD-based wind engineering applications, particularly in urban aerodynamics, structural wind loading, and environmental flow modeling. The study also highlights optimal mesh thresholds beyond which computational cost outweighs accuracy gains, offering practical guidance for high-fidelity numerical simulations.