High-resolution digital image correlation (HRDIC) is a technique used to study the deformation of metallic materials. By combining HRDIC measurements of local strain with orientation information from electron back-scatter diffraction, it is possible to determine strain localization, slip activity and relate it to the underlying microstructure. These results can then be compared to the results of crystal plasticity full field modelling (CP), for either calibration or validation. However, because HRDIC measures only the strain at the surface, and EBSD only provides access to the orientations and partial shapes of the grains at surface, direct comparisons are not possible, since the effect of subsurface microstructure on the surface deformation is unknown. In this work, a stochastic Monte Carlo Markov Chains (MCMC)-like approach using centroidal Voronoi tessellations has been employed to generate representative volume elements (RVE) for CP-FFT, in which the surface microstructure is invariant, but the subsurface microstructure is varied randomly. Grain shapes and positions can be changed by perturbing the seed positions, and the subsurface orientations can be shuffled to modify the misorientation distribution, to evaluate the effect of texture clustering on the surface response, for example. Using an iron material model, we further investigated how the subsurface morphology and texture affects the slip system activity, texture evolution and strain localisation in body-centered cubic metals. Our results showed that, as previously reported, the subsurface microstructure greatly affects deformation evolution on the surface, with the grain morphology having the greatest effect. However, the overall slip activity is not significantly affected, implying it can be reliably used for CP model calibration. Nevertheless, for a meaningful comparison between HRDIC and CP-FFT, the effects of the subsurface should be considered by comparing the experimental results to the distribution of predictions from the simulations of several equivalent but different volume elements.