<p>Foundations are fundamental components of most modern above-ground buildings and structures, but are problematic for redevelopment on cleared brownfield sites since they potentially remain buried and un-identified, and could negatively impact on future construction work. Prior to redevelopment of such sites there is therefore a need to both detect and characterise such remaining features, ideally initially using non-invasive investigation methods to identify areas of interest that can then be investigated by more expensive, intrusive methods. Electrical resistivity surveys are one of the geophysical methods that have been used for site investigations. They are commonly acquired using two-dimensional (2D) surveys using specific array types which give differing results. This paper aims to show how different array types change the electrical resistivity responses of common shallow foundation types using 2D numerical modelling methods. Electrical resistivity datasets were synthetically generated and inverted for 5 common foundation styles using five common resistivity array types. Results show that the dipole–dipole array was optimal in identifying the base of isolated simple and stepped foundation footings, whereas the Schlumberger, dipole–dipole and pole–dipole resistivity arrays were optimal for more complicated foundations. For piling foundations it was determined that pole-dipole and pole-pole arrays were optimal. This study suggests using numerical electrical resistivity modelling of foundation targets to optimise geophysical surveys before conducting brownfield site geophysical surveys.</p>