The evolution of decision support systems (DSS) has generally been in pace with the advancement of computing and communication technologies. More recently, there has been an enormous enhancement of computing power and internet bandwidth, coinciding with a proliferation of web-based technologies. This has opened up new opportunities in terms of using complex modelling for flood forecasting and communicating the projections. However, a literature review shows that many of the contemporary DSS in the field of flood emergency management employ surrogate or simplified flood modelling systems. The reason is that undertaking complex flood simulation models has been considered as infeasible in the short time available during a flood emergency. This paper challenges that paradigm. The desire for this paradigm shift is underpinned by the recent advent of Graphic Processing Unit (GPU) flood modelling systems and sophisticated web-based GIS systems that can better present the results of these models. Furthermore, it is proposed that there should be a move away from model-based systems to open systems that can house modelling engines and communicate the outputs effectively for decision-makers. This approach promotes user-focused communities that can cross agency and proprietary lines and reduce costs and promote the maintenance of developed systems. Emergency management decision-making is usually threshold based, and accordingly, outputs should be produced that facilitate this type of decision-making. This paper describes the framework and the working prototype of a flood emergency decision support system (as a proof of concept) that is in operation and supports tactical and strategic decision-making during flooding on the Gold Coast.