Estimation of climate change impacts on river flow and catchment hydrological connectivity incorporating uncertainty from multiple climate models, stochastic downscaling and hydrological model parameterisation error sources

When estimating climate change impacts, there are many sources of uncertainty which must be considered. The main sources of uncertainty arise from the structure and parameterisation of physically based simulation models, downscaling methods, stochastic realisations of future weather time series and the underlying emission scenarios. This work focuses on the uncertainties resulting from the use of multiple climate models and the joint impact of the stochastic realisations of future weather time series from a weather generator, EARWIG, and from parameter estimation uncertainty of a hydrological model, CAS-Hydro. These tools have been applied to the River Rye, Yorkshire. A suite of model parameter sets and weather realisations have been used to project likely changes to the hydrological functioning under climate change. Results are presented on the projected changes in flow duration curves and the potential changes in the hydrological connectivity by overland flow within the catchment. The statistical sensitivity of the impact predictions to these sources of uncertainty and the use of a multi-model ensemble to enable the production of probabilistic estimates of change is assessed. These estimates of potential changes in flow can then be used to inform the adaptation of water resources design and management.