While initial studies have demonstrated the feasibility of applying two-dimensional finite element models to reach scale river flood flows, the further development and practical use of such codes for engineering problems has been hampered by a lack of available validation data. This paper attempts to address this research need by conducting benchmark model validation studies over a range of reach scales, flow regimes and flood events. Specifically, four river reaches in the UK (the Culm, Stour, Severn and Blackwater) and one in the USA (the Missouri) have been modelled using the TELEMAC two-dimensional finite element scheme. These simulations have been compared in terms of the ability of the model to simulate both flood routing and stage hydrograph data at points internal to the modelled domain. This has enabled a more critical assessment of model capabilities to be obtained and has allowed the better definition of future research needs. In particular this study has identified the need to employ data internal to the model computational domain to achieve robust validation and the need to now undertake comparison of two-dimensional schemes with alternative flood simulation tools.