A history of Emergency Alerts in the UK. Part 1: The early years (2012–2018)
The UK’s journey with Emergency Alerts has certainly not been straightforward. Despite being a slog at times, I’ve been fortunate enough to have been involved almost all the way through its development.
As I write this the Emergency Alerts service for the UK has yet to go live. When I departed the national project team in April 2021, the plan was to launch in June that year. However, there are signs the launch may be back on the table with an awareness campaign together with a national ‘welcome message’ that could be sent in the near future [now planned for the 23rd April 2023 at 3pm].
I’ve split this blog into three parts and I’ve not attempted to cover every aspect, after all this history spans more than 10 years! I‘ve focused on the highlights and mentioned those who’ve helped shape the UK’s public warning system.
The National alerting capability
Cabinet Office Civil Contingencies Secretariat (CCS) started to explore the introduction of a national alerting capability in 2012. This was prompted by the Strategic Defence and Security Review from 2010 which recommended an evaluation of options for an improved alerting system.
This initially started with limited testing of existing technology such as the Environment Agency’s (EA) flood warning service which, at the time was limited to sending alerts via phone, fax or pager on an opt-in basis. The scenario that was selected was warning for an industrial accident at a site in Ellesmere Port. Perhaps unsurprisingly the feedback from those involved in the exercise recognised the need to warn people based on their location and through enrolment rather than opting in which traditionally has low rates of take up.
The trial at Ellesmere Port in conjunction with the EA and Cheshire Local Resilience Forum (LRF), led to a subsequent and more substantial trial which became known as the Mobile Alerting Trials (MAT). Again led by Cabinet Office CCS, much of the groundwork was laid during this time through the dedication of two individuals, David Barnes and Charlotte Lawrence. The problems the Cabinet Office was seeking to address is summarised in this graphic.
This led to a number of work streams taken forward by the CCS team during the period 2013–2015 which included:
- Updating the Privacy & Electronic Communications Regulations (PECR) to include the provision of sending alerts to citizens as a result of emergencies.
- Initiating a standards challenge regarding the introduction of a standard for the exchange of Emergency Messages*
- Developing templates and protocols for the type of incidents that would make use of the technology
- Delivering a trial of location based SMS (LBSMS) and cell broadcast (CB) technology (Mobile Alerting Trials).
*The original challenge proposing the use of the Common Alerting Protocol (CAP) was led by the Met Office but didn’t complete. In 2021 Simon Nebesnuick, Ray Hanson and I revisited this to progress this to adoption. You can read more about the benefits of CAP as a standard here.
LBSMS & Cell Broadcast (CB)
The two main direct alerting technologies are LBSMS and CB and whilst mobile apps have also been used in some countries, the take up rate has been poor and reliance on opt-in methods have generally failed to result in any significant reach.
As Cell Broadcast is less well known in the UK, here’s a summary of the three main components that make up a CB-based warning service. Firstly a message sending platform is required known as a Cell Broadcast Entity or CBE (1). This is where the authority sending the message defines the message content and location in which recipients are intended to receive the alert.
Once issued the alert is passed through to the Cell Broadcast Centre or CBC (2). The CBC is software that sits inside the mobile operators network so for the UK there are four for each of the 4G/5G networks with an additional CBC for contingency. The CBC matches the location specified from the CBE to the mast infrastructure in that area. The CBC also confirms the start and the end time of the broadcast which can be close to, but not exceeding, 24 hours. The repetition rate is also confirmed, this being the rate at which the alert is broadcast e.g. every 120 seconds. If you were outside the alert area but travelling into that area you could expect to receive the alert within that time frame.
Finally we have the handsets (3) which require the right software for the cell broadcasts to be received and displayed correctly. Many handsets are ready ‘out of the box’ but in the case of Apple and the main Android manufacturers an over the air install is required to activate this and to ensure the correct spec defined for that country is displayed (more on this later).
Both LBSMS and Cell Broadcast have pros and cons summarised below.
Whilst this is the current picture, much the development of cell broadcast has occurred in the last five years. The potential for CB is huge, particularly with 5G which opens up the potential to provide maps and icons with the Alerts.
It is a very different proposition now compared with the technology that was only available on the 2G network where phones were highly inconsistent with how alerts were presented. In effect they appeared like sms messages.
The public is also far more conscious of privacy and protecting their personal data. Both the original lockdown SMS in March 2020 and more recently with the ‘Get Boosted’ SMS in December 2021 have highlighted how strongly people feel about authorities using their contact details regardless of the scenario. With CB not requiring personal information and being one-way, this is a far more palatable proposition for those whom privacy is a genuine concern.
Mobile Alerting Trials (MAT)
The MAT set out to technically test LBSMS and Cell Broadcast together with capturing views on the introduction of such a service from both the responder community and public. The full report is still available for those who want to delve into more detail but here’s the shorter version.
Three locations for the trials were selected, these were Glasgow (LBSMS), the Suffolk Coast (LBSMS) and Easingwold, North Yorkshire (CB). The trials were conducted independently a few months apart with efforts placed into communications to raise awareness ahead of each trial.
Glasgow was selected for the simulation of a evacuation as a result of an explosion within the city centre. O2 supported this test with 25,000 messages sent to 11,000 recipients. Suffolk was chosen for a civil nuclear infrastructure, a substance release from Sizewell B. This trial was supported by Vodafone with 11,000 SMS messages sent. The North Yorkshire trial was supported by EE and simulated a severe weather event targeting 50 recipients. This rural scenario enabled a cell broadcast, untested in the UK at the time, to be reasonably contained. Three UK declined to take part in the trials.
The Cell Broadcast Test in North Yorkshire was conducted at the Emergency Planning College at Easingwold. The Netherlands had recently gone live with the introduction of their nationwide alerting capability on the 2G network. For the trial EE sought support from operations in the Netherlands to enable the limited UK test. A single mast was chosen and an additional test with SMS was conducted as a comparison for the 26 supplied supplied handsets.
Volunteers based at the Emergency Planning College were asked to note down the timing and type of messages received.
The MAT concluded that LBSMS was a better placed technology to provide a national emergency capability which was an understandable conclusion given CB technology was very early in its development at this point. Many of the limitations we experienced at Easingwold with CB have now been overcome as the tech has matured substantially;
- We have seen the dominance of Apple, Google and Samsung. The consolidation of operating systems (iOS and Android) have enabled a consistent user experience which was one of the key concerns from the trials.
- Standards have also developed, further ensuring consistency globally but encouraging further developments such as extending the maximum character length of messages from just 93 characters, as was experienced in 2013, to 1395 (15 pages of text), additional language support and with WEA3.0 the introduction of more targeted messaging through device based geo-fencing (DBGF).
- In 2013, some phones required users to opt-in to receive alerts by adding specific channel numbers into the handset settings. In 2022, whilst some phones will still allow users to add legacy channels, the adoption of newer standards (3GPP) and the uptake of newer devices has eliminated the need for users to opt in.
- Importantly the user experience has transformed from a rather passive presentation of an alert similar to that of a text message to an alert which looks and feels very different with a distinctive tone, vibration and the ability to override volume settings.
Since 2013 we have seen the adoption of cell broadcasting in many other countries such as New Zealand, Canada, Greece, South Korea, the UAE to name several. Implementation time and costs have also reduced considerably. The most notable being Romania which implemented their national emergency alert service in less than 4 months at a cost of €3million euro.
The CCS team continued to press for the development of the national alerting capability, creating Alert Referencing Groups to create and design templates for message content and agree the scenarios under which the service would operate. User stories were also produced as the team focused on the needs of the service users, both those initiating the messages and those receiving them.
Progress stalls
CCS planned to take a paper to the Cabinet Committee for flooding in 2014 to gauge views on taking the work forward based on the understanding of the expected costs at that time. Securing budget for this work was always problematic. Various departments were approached to contribute to the service on the basis they would be users but this failed to gain any traction.
Emergency Alerting has always fallen between two departments. From a national emergencies perspective this is the lead of the Cabinet Office Civil Contingencies Secretariat (CCS), a relatively small unit within Government. From a telecommunications perspective the lead is the Department for Digital, Culture, Media and Sport (DCMS).
With one department not owning the whole thing, prioritisation of this work was always going to struggle. 2015 then brought an election which diverted attention and in 2016 the Brexit Referendum derailed any plans of taking forward the implementation of a national alerting capability.
Political pressure grows
Whilst there was no outwardly visible progress from Government around this time, there was still political pressure and further high profile incidents which put a spotlight on the need for such a service. In 2016 Lord Toby Harris published the independent review for the Mayor of London on London’s Preparedness to Respond to a Major Terrorist Incident, this included a specific recommendation (110):
The Mayor should quickly work with the Cabinet Office to introduce a London-wide pilot of public alert technology.
In 2017 The City also experienced the London Bridge Attack and the Grenfell Tower fire, both of these reconfirming the need for an emergency alerting capability that would have undoubtedly saved lives. In the case of Grenfell, when the advice shifted from ‘stay put’ to evacuate, the authorities had no reliable way of alerting those in the tower to this change of status.
At CCS a new lead for this work was appointed, Dr Nigel Brown. Nigel worked across a number of projects in civil contingencies. In this article from 2009 Nigel is speaking about the impact on the resilience of communications in a flu pandemic. In 2018, CCS procured an Insight Report produced by Fujitsu to assess the alerting landscape both domestically and internationally given it was now some five years since the original Mobile Alerting Trials.
2018 also saw the introduction of new European Electronic Communications Code (EECC) Directive (Article 110). This Directive required that:
By 21 June 2022, Member States shall ensure that, when public warning systems regarding imminent or developing major emergencies and disasters are in place, public warnings are transmitted by providers of mobile number-based interpersonal communications services to the end-users concerned.
Whilst ultimately the UK left the EU, this Directive prompted a number of countries to plan and implement national alerting capabilities. The number of commercial operators in this space is relatively small and many took the opportunity to suggest countries must implement an emergency alert capability requiring the use LBSMS or a combination of CB & LBSMS due to the requirement for warnings to be transmitted to “all” end-users concerned. As Cell Broadcast is one-way and is privacy free without the need to know the recipients phone number, it is not possible to confirm receipt.
The Body of European Regulators for Electronic Communications (BEREC) helpfully produced guidelines to provide clarity:
Because of the sentence “when public warning systems [..] are in place”, Article 110 of the EECC does not place any obligation upon Member States without existing PWS, to develop or deploy a legacy or ECS-PWS.
Furthermore, in relation to transmitting to to “all” end-users concerned. BEREC concluded that there may be practical reasons why this requirement may not be achievable for example temporary terrain shielding of signals due to cranes or other temporary obstructions in the vicinity of warning transmission or the end-user could be in a basement without reception or their device could be out of power. BEREC therefore encouraged public warning systems to reach as many concerned end-users as technically possible.
Much of the available literature relating to LBSMS at this time was based on the implementation of an Emergency Alert service in Australia (2011). By this point the technology was over 7 years old, a lifetime in digital and technology. The suggestions that CB was an inferior platform seemed at odds with the development of the technology and the increasing number of international deployments.
Mobile UK, the trade association for the UK’s mobile network operators supported the view that CB provided the best option to deliver the requirements of Article 110:
“Having considered the matter recently, in light of the requirement in the EECC, the Mobile Network Operators feel that Cell Broadcast will be much better at fulfilling the requirements to get a public warning alert delivered in a timely manner. Subject to a number of factors (such as the size of the defined area, the number of people to be contacted etc.), delivering, 100,000 messages, say, by SMS could just take too long to be of any practical use in an emergency situation.”
Having worked in incident management for over 18 years by this point I wanted to understand the potential for CB in the UK, outside of continual references to the original Mobile Alerting Trials which was a very limited 2G network test when the technology was in its infancy. This interest lead to a further trial of Cell Broadcast which I’ll cover in Part 2.
