In systems of interconnected ponds or lakes, the dispersal of zooplankton may be mediated by the active population component, with rivulets and overflows functioning as dispersal pathways. Using a landscape-based approach, we modelled the effective geographical distance among a set of interconnected ponds (De Maten, Genk, Belgium) in a Geographic Information System (GIS) environment. The first model (the Landscape Model; LM) corrects for the presence of direct connections among ponds and was based on the existing landscape structure (i.e. network of connecting elements among ponds, travelling distance and direction of the current). A second model (the Flow Rate Model; FRM) also incorporated field data on flow rates in the connecting elements as the driving force for the passive dispersal of the active zooplankton population component. Finally, the third model (the Dispersal Rate Model; DRM) incorporated field data on zooplankton dispersal rates. An analysis of the pattern of genetic differentiation among Daphnia ambigua populations inhabiting 10 ponds in the pond complex reveals that the effective geographical distance as modelled by the flow rate and the dispersal rate model provide a better approximation of the true rates of genetic exchange among populations than mere Euclidean geographical distances or the landscape model that takes into account solely the presence of physical connections.