Evolution of flow characteristics through finite-sized wind farms and the influence of the wind-farm configuration on modulating this evolution is explored through numerical simulations. The principal aim for the study is to identify regions of flow-adjustment and flow equilibrium within the wind farm. Towards this aim, a suite of five large-eddy simulations (LES) of the neutral atmospheric boundary layer with extremely long streamwise domains are performed with embedded finite-sized wind farms of different streamwise and spanwise spacing. Three diagnostic variables, namely, the wind-farm induced effective surface roughness, the wake viscosity and the wake-expansion coefficient are computed using the LES-generated database and are used to characterize the flow. Computation of the diagnostic variables is relevant to the wind-energy community in different contexts ranging from parametrization of wind farms in weather and climate models, to wind-farm design and optimization based on wake models and eddy-viscosity type Reynolds-averaged Navier-Stokes solvers. Results show that flow equilibrium is achieved in the 'most dense' configuration of s(x) approximate to 8D, s(y) approximate to 5D at approximately the 19th row. Results also indicate that the streamwise spacing plays a dominant role determining the rate at which flow-adjustment is achieved within the wind farm. (C) 2017 Elsevier Ltd. All rights reserved.