The development of earthquake early warning systems and methodologies in the last two decades resulted in one of the most important contributions of the seismological and engineering communities to risk mitigation. However, the great variability in economic resources, scientific background and technical preparation between developed and developing countries makes it indispensible to further develop approaches that can be adapted to these very different situations. Furthermore, several prominent recent natural disasters have shown that future systems should be able to analyse in real time and provide rapid information and alarms for a range of different natural disasters (both when a triggering mechanism exists or not). The aim of this session is to provide a state-of-the-art in the field of early warning systems, with a particular, but not exclusive, attention on earthquakes, tsunamis and landslides. We welcome contributions on early warning developments, both on algorithm as application level, and particularly encourage submissions dealing with multi-risk applications and loss-based early warning.