Inter-catchment groundwater flow (IGF) can affect the quantity and quality of freshwater as it connects topographically separated catchments. Therefore, assessing the space–time variability of IGF is important in water resources management. Traditional measurements of rainfall, evapotranspiration, and runoff can provide indications of IGF through catchment-scale mass balance analysis. However, it is difficult to study the detailed space–time variability of IGF using measurements only. Current rainfall-runoff models often either neglect IGF completely or include an over-simplified representation disregarding the spatial variability of IGF. Three-dimensional groundwater flow models can simulate the space–time variability of IGF, although they require a substantial amount of data and computational resources. In comparison, a two-dimensional model that simulates groundwater flow only horizontally considering depth-averaged hydrogeological properties, demands significantly less subsurface data and computational resources. This study evaluates the performance of a two-dimensional groundwater flow model in simulating IGF across the topographic boundaries of two chalk catchments, namely Kennet (1033 km²) and Pang (171 km²) located in the South of England. Using publicly available data, our proposed model is capable of reasonably reproducing the variability in observed flow – with Kling-Gupta Efficiencies of 0.91 for Kennet and 0.8 for Pang – as well as groundwater head, achieving a coefficient of determination of 0.85. We demonstrate that our model simulates the space–time variability of IGF and closes the mass balance within the topographic catchment boundaries consistent with observations. We find that a topographic catchment can lose 19 % of its annual precipitation as IGF, supporting the needs for including the space–time variability of IGF in freshwater management.