In Australia, the best use of limited national water resources continues to be a major political and scientific issue. Average water allocations for rice-cereal irrigation farmers in the Riverina region have been drastically reduced since 1998 as a consequence of high rainfall variability and prolonged periods of drought, together with political changes. This has severely impacted regional crop production during the last decade, threatening the livelihoods of many farmers and is in stark contrast to much of this region’s 100 year agricultural history, where water resources were available to farmers in steady abundance. The water ‘landscape’ has changed - bringing with it considerable social, economic and environmental consequences and forcing a rethink of how valuable water resources are best used under such variable, changed and changing conditions. This thesis presents details of investigations into on-farm adaptation options for rice-cereal farmers, using field experimentation, participatory engagement, and farming systems modelling as the major tools of research. Additionally, a major component of this work has been the development and testing of new modelling tools and decision-support structures. Well-tested cropping systems models that capture interactions between soil water and nutrient dynamics, crop growth, climate and management can assist in the evaluation of new agricultural practices. At the beginning of this research project, all available models were lacking in at least some major element required for simulation of rice-based cropping systems. The capacity to simulate C and N dynamics during transitions between aerobic and anaerobic soil environments was added into the APSIM model, to facilitate our need to model farming system scenarios which involved flooded rice in rotation with other crops and pastures. Thorough testing against international datasets was subsequently conducted. Photosynthetic aquatic biomass (PAB – algae) is a significant source of organic carbon (C) in rice-based cropping systems. A portion of PAB is capable of fixing nitrogen (N), and is hence also a source of N for crops. To account for this phenomenon in long term simulation studies of rice-based cropping systems, the APSIM model was modified to include new descriptions of biological and chemical processes responsible for loss and gain of C and N in rice floodwater. Using this improved APSIM model as a tool, together with participatory involvement of Riverina case-study farmers, it was demonstrated that the best on-farm cropping and irrigation strategies in years of high water availability were substantially different to those when water supplies were low. The strategies leading to greatest farm returns vary on a season-by-season basis, depending primarily on the water availability level. Significant improvements in average farm profits are possible by modifying irrigation strategies on a season-by-season basis. The opportunities for Riverina farmers to exploit their irrigation water resources also extend beyond the farm gate. Currently there is considerable confusion amongst farmers on how to evaluate and compare on-farm and off-farm water options. Direct selling of water seasonally on the open market and even permanent sale of irrigation water entitlements are possibilities. In response to this confusion, a new conceptual framework was developed that enables quantitative comparisons between various options. The framework is based on a method regularly employed in the financial world for share portfolio analysis. Simulation modelling provided risk-return characteristics for on-farm options, and helped to elucidate circumstances under which off-farm options were viable. A modified version of alternate wet-and-dry water management for Australian rice-growing conditions (delayed continuous flooding, DCF) was investigated via a 2 year field experiment – aimed at reducing irrigation water requirement and increasing water productivity (WP). We demonstrated up to a 17% increase in WP, and field data was generated on system performance for a range of discrete irrigation strategies. The APSIM model was then parameterized, calibrated and validated before being used to extrapolate findings from the two year experimental period to a much broader climatic record (55 years), allowing detailed investigation of optimal management strategies and a more realistic estimation of likely long-term gains in water productivity, and associated risks, from this new rice irrigation practice. Best practice guidelines were developed, and the potential impact of a changing climate on both optimal practice and likely benefits was assessed. This thesis concludes by synthesising the approaches taken - addressing the question of whether improved rice irrigation practices, seasonally-flexible cropping and irrigation strategies and off-farm exploitation options, can in combination address the challenges of reduced irrigation water allocations in Australia’s Riverina region. Evidence is presented that the answer is yes under certain circumstances, but that limits to change exist beyond which the investigated on-farm adaptations are not enough. The thesis also proposes that the concepts and methods developed during this project are globally applicable and useful in the design of farming system adaptation options. Keywords: irrigation, limited water resources, farming systems modelling, participatory engagement.