This paper describes a software modelling system, SimCLIM, for simulating the effects of climate variability and change, and demonstrates its application to domestic rainwater harvesting (RWH). The paper first describes SimCLIM and then focuses on a case study of the reliability of domestic RWH systems as the sole source of water in South East Queensland, a region of high spatial and temporal variability in rainfall. Using the SimCLIM scenario generator along with gridded daily time-series data and a domestic water tank model, 'risk landscapes' are created based on the risk of failure of the domestic systems to provide water under both current climate and scenarios of future change. Steep risk gradients were found, with major implications for RWH system configuration. Simulations are then performed in order to examine the trade-offs between two broad strategies - increasing runoff area and increasing tank size - required to maintain a reliable system. The paper concludes that, instead of set specifications, location-specific customisation of RWH systems is preferable in order to ensure reliability under current rainfall variability and sustainability over the long term under climate change. The availability of gridded daily rainfall datasets along with modelling tools like SimCLIM makes this customisation practicable.