Hurricanes and tropical storms occur very frequently in the Central American and Caribbean Region (CA&CR). These extreme weather events produce a lot of rain and consequently a lot of flooding. Damages and loses have been estimated to amount to 13.4 billion dollars in the last ten years. Flood Warning Systems (FEWS) are a key preventive strategy to reduce risk. Technological progress is improving the resources made available for FEWS to be more viable. In spite of the international support for FEWS and the fast development of ICT, there are very few countries in the CA&CR that have succeeded in developing fully operational warning systems that are functioning in a sustainable manner for a long period of time. There is disconnection between the community-based systems and the centralized systems, as well as between the National Meteorological Services (NMS) and the National Hydrological Services (NHS) which tend to work many times in isolation. The general purpose of this work is to unravel the dysfunction/chaos of the way early warning systems are done and provide guidelines to integrate flood warning system at all scales to be used in operational forecasting, particularly for countries in the CA&CR. Flood warning can be seen as a set of sub-systems in which forecasting is only one of those sub-systems. A conceptual framework has been proposed to classify flood warning systems using the spatial and temporal scale at which the flood warning systems operate, subdividing them into Global, Regional, National and Local FEWS. In practice, these systems are not operated in an integrated manner. Emerging technology is available to allow the integration of global- and local-scale forecasting resources in the CA&CR. The Tethys Platform has a series of online tools applicable to flood forecasting. A workflow is given for the use of four apps in Tethys for flood forecasting: (i) Stream flow Prediction Tool; (ii) Reservoir operation tool; (iii) Hydro Viewer Hispaniola; Flood Map Visualization tool.