Abstract Volcanic eruptions pose a significant risk to human lives, property and infrastructure, despite rapid advances in monitoring and early warning science and technology. Some elements of risk – such as the number of people living close to volcanoes – are increasing, and the unpredictable nature of eruptions may overwhelm the local response capacity and turn into a disaster, sometimes requiring international assistance. To deal effectively with these crises, the international humanitarian community needs a global, science-based early warning system that should assimilate the state-of-the-art monitoring and early warning techniques, as well as being able to provide a preliminary impact assessment, and issue appropriate and relevant alerts. Current volcano warning systems are either only local in context or are not suited to the needs of global early warning. In this paper we propose an outline for a volcano warning system aimed at issuing alerts to the humanitarian aid community. It is designed as a four-level system, incorporating the latest monitoring and hazard modelling techniques that are applicable on a global scale. Alerts are mainly based on the predicted humanitarian impact of the modelled hazards. Systematic handling of volcanic manifestations, such as thermal signals and ash clouds from space-borne instruments, make it possible to create such a system. The Global Disaster Alert and Coordination System (GDACS), a joint effort by the United Nations and the European Commission, has been operating in a similar spirit for other natural disasters for a number of years and could fulfil the role of the desired volcano system. This paper discusses the needs and issues of this undertaking.

Abstract Using two hypothetical effusive events in the Chaîne des Puys (Auvergne, France), we tested two geographical information systems (GISs) set up to allow loss assessment during an effusive crisis. The first was a local system that drew on all immediately available data for population, land use, communications, utility and building type. The second was an experimental add-on to the Global Disaster Alert and Coordination System (GDACS) global warning system maintained by the Joint Research Centre (JRC) that draws information from open-access global data. After defining lava-flow model source terms (vent location, effusion rate, lava chemistry, temperature, crystallinity and vesicularity), we ran all available lava-flow emplacement models to produce a projection for the likelihood of impact for all pixels within the GIS. Next, inundation maps and damage reports for impacted zones were produced, with those produced by both the local system and by GDACS being in good agreement. The exercise identified several shortcomings of the systems, but also indicated that the generation of a GDACS-type global response system for effusive crises that uses rapid-response model projections for lava inundation driven by real-time satellite hotspot detection – and open-access datasets – is within the current capabilities of the community.