Future Water Vulnerability in Ireland: An Integrated Water Resources, Climate and Land Use Changes Model

Water resources management and policies need to consider the dynamic nature of any catchment’s water balance, particularly in planning stage, to develop effective strategies for the future. The main goal of this research is to create an innovative and integrated environmental modelling tool (GEOCWB) by applying Machine Learning Techniques to a Geographic Information system (GIS). The developed tool uses as test and validation case the trans-boundary Shannon river basin. Climate change projections for the Shannon River catchment are simulated and presented using GEO-CWB for several climate variables from multi-GCM ensembles for three future time intervals using a range of different Representative Concentration Pathways (RCPs). As part of the integrated environmental modelling approach, the future spatially distributed urban expansion scenarios and land use changes for Shannon river basin are simulated and presented based on realistic land cover change models and projected to several time intervals. This is achieved using a hybrid modelling technique combining a logistic regression and a cellular automata (CA) model for developing spatial patterns of urban expansion. The research presented here provides an appropriate methodology for long-term changes analysis in European trans-boundary river’s water level and streamflow parameters after using a customized GISbased algorithm to simplify the hydrological system. GEO-CWB provides an integrated GIS tool for modelling potential evapotranspiration on the catchment scale. The GEO-CWB tool has been developed to help and support water sector modellers, planners, and decision makers to simulate and predict future spatially distributed dynamic water balances using a GIS environment at a catchment scale in response to the future change in climate variables and land use. Several Machine Learning Techniques are applied on the outcomes of the GEO-CWB model for the Shannon River in order to model and predict the water level and streamflow parameters for some stations along the river for daily time steps.