This thesis presents an analysis of uncertainty at the local stage of a climate impact assessment. Impact model structural uncertainty and uncertainty due to equifinality of parameter sets are evaluated, in addition to uncertainty due to GCMs and emissions scenarios. The Suir catchment is employed as a case study area to analyse the changes in catchment hydrology and in future flood magnitude and frequency relationships due to climate change. Two lumped conceptual rainfall-runoff models of different degrees of complexity are forced with the output of three GCMs and two emissions scenarios (A2 and B2) downscaled to synoptic station level by empirical statistical downscaling (Fealy and Sweeney, 2007). In the analysis of changes to catchment hydrology for the 2050s and the 2080s, GCM uncertainty is the greatest source of uncertainty. However, by the 2080s, uncertainty due to equifinality of parameter sets and model structure is also a significant source of uncertainty, with increases in streamflow being most extreme in February. Furthermore, results suggest that flood magnitude and frequency relationships will intensify under climate change. A robust finding is the notable agreement in new return period values in the 2080s with both models suggesting that the 10, 25 and 50 year flood events simulated in the control period will become 3.2, 5.4 and 9 year flood events. However, the magnitudes of the flood events differ for each model. These results suggest that model structural uncertainty is a significant source of uncertainty and should be taken into account by employing a suite of hydrological models at the local stage of climate change impact analyses that inform anticipatory flood adaptation decisions or policy frameworks.