The mid-Piacenzian warm period (3.264Ma-3.025Ma) was the most recent interval of Earth's history that was distinctly warmer than the present for a sustained period of time. As it is geologically recent, it shares many similarities with the modem world such as continental locations, topography and currently extant species of flora and fauna. Proxy data have suggested that the atmospheric concentration of carbon dioxide was around 400ppm, very close to the modem level. These factors taken together have made the mid-Piacenzian attractive to many as offering a potential analogue for the climate of the coming century, both in terms of gaining insights into a warm climate and as a testbed for the climate models that are also used to make projections of future climate change. Atmosphere only and coupled atmosphere-ocean simulations of the mid-Piacenzian climate have been produced using the Hadley Centre models (HadAM3 and HadCM3) that are compliant with the framework of the Pliocene Model Intercomparison Project (PlioMIP). These simulations constitute the UK contribution to the PlioMIP ensemble. In view of the ongoing efforts to constrain proxy reconstructions of mid-Piacenzian atmospheric carbon dioxide concentration, an ensemble of simulations based on the coupled HadCM3 simulation has been generated for a range of CO₂concentrations in order to assess the response of the climate to uncertainty in this key boundary condition. A spatial statistical model-data comparison method has been developed that is able to take the scattered data observations and create an objective skill score that can be used to rank model performance. This method is first tested against pseudo-observations derived from the CO2 ensemble to test its ability to select the correct model. Finally, it has been used to rank the CO2 ensemble against a global dataset of mid-Piacenzian sea surface temperature data. The PlioMIP simulations capture levels of warming seen in the observational data, though the coupled model fails to capture the full extent of polar amplification. The results fall within the range of climates from the wider PlioMIP ensemble. The model-data comparison method is able to select pseudo-observations from the CO₂ ensemble successfully when control parameters are chosen that are suitable for the distribution of observations. A global assessment of the data against the CO₂ ensemble favours the PlioMIP standard CO₂ level of 405ppm. A more detailed regional assessment suggests that 350-405ppm is the likely range of mid-Piacenzian CO₂ concentration