Detection, Alert and Warning Systems (DAWS)
Design
By Dr Peter Williams chairs ARISE-US. His background includes 30 years in IBM, where he became an IBM Distinguished Engineer, and extensive experience in creating DRR tools such as the UN City Disaster Resilience Scorecard and its many offshoots, now used by hundreds of cities (and countries) globally. His PhD is in Politics.
With the launch in 2022 of the UN Secretary General's Early Warnings for All initiative, the provision of warning systems is deservedly attracting great worldwide attention. Yet the emergence of criticisms of the performance of alerting systems in incidents ranging from floods in Spain and Australia, to wildfires and tsunamis in the Western US and Hawaii suggests that even where such systems exist, shortcomings may still arise.
An alert is not just an event - it is a process, spanning event detection through delivery to action being taken by the affected population. That process may malfunction, or at least lack adequacy, at multiple points. As examples:
Either through insufficient sensitivity or through malfunction, the detection system may fail to notice the event. Alternatively it may falsely report an event that does not exist, thus undermining confidence in it.
The detection system may not respond in sufficient time to take action, either due to the inherent nature of the hazard (earthquakes, flash floods, tornadoes) or because transmission of data to the warning system may be delayed by lack of bandwidth or other technical issues.
Translation of data into appropriate warnings and the distribution of these may also be delayed for bureaucratic reasons - failure to act fast enough, or to coordinate between multiple agencies.
The alerting system may lack multi-hazard capability - the ability to differentiate between different disaster events, resulting potentially in wrong actions being taken. The ability to deal with cascading hazards (for example an earthquake followed by a dam failure) is a subset of this problem.
The alerting system may issue warnings that are too broad, for example where people receive evacuation notices when they are far enough from the event or its impact that they do not need to evacuate. The warnings may lack context - for example, graduations to denote "be ready" and then "evacuate now".
Different agencies and jurisdictions affected by the event may not have coordinated alerting thresholds, responsibilities and processes for the area in question - resulting in inconsistencies and contradictions between different parts of the affected area, or for different systems (for example energy, water, transportation) within the area.
Critically, the alert or warning may not reach all of the affected population. Not everyone has a smartphone!
Alerts and warnings may not be delivered in a format, and with whatever supporting information may be required, to enable rapid, appropriate action by different groups of recipients (for example linguistic or social groups) from different parts of the event area. Alerts and warnings may not be accessible – over 1.3 billion people across the globe have disabilities, and many more experience low or no literacy.
Where they exist, DAWS may be obsolete due to under-use of the technologies that are available, for reasons of cost, lack of awareness, trust or inertia.
DAWS may not be integrated with critical systems - PA systems, radio, TV, automated shut-down systems (eg for metros, safety-critical industrial processes) and so on.
It is clear that the "Early Warnings for All" initiative needs to take the entire detection alert and warning process into account - not just the warning device itself. ARISE-US has commenced an investigation in to the potential shortcomings in the process with a view to identifying what works and what does not in different contexts globally. We will begin with symposium to explore the issues and continue with research and policy proposals to address these. We look forward to publishing our results in due course.
