Emergency Alerts & Storm Babet

Why was the new Emergency Alert Service not used in the recent floods?

I’ve been messaged several times this week asking this question. Whilst it’s been over two and half years since I worked on the service, no-one else from the national team seems to written anything, at least publicly, about the service leaving lots of questions unanswered.

To answer this question I need to provide some context. The original use case was for Covid-19. It was this that resurrected the idea of a national alerting capability after the difficulties experienced sending national SMS messages which required batching to avoid network congestion resulting in a large number recipients receiving the messages over 48 hours later.

I joined the national emergency alerts team on loan from the Environment Agency and was keen to ensure flood risk was kept high on the list as a use case. As the threat from Covid started to reduce, the likelihood of flood risk being a candidate for use, started to increase.

National Protocol

Work on Emergency Alerts was split across various departments with DCMS (Department for Culture, Media & Sport), leading on the overall delivery with the expectation of a handover to the Cabinet Office on the completion of the project. Cabinet Office led, as you might expect, on the National Protocol — essentially the document which described who could use the service, how the service should be used and what it shouldn’t be used for.

Cabinet Office also brought together various individuals to form an Oversight Board. In this screenshot i’m presenting the potential use cases and their priority at that time.

Just to be clear, it was never the intention to launch the service covering all of these use cases from day one. The system would be iterated to meet the needs of the various agencies through their onboarding.

It was however difficult to engage the Devolved Administrations. Reaching the right people and getting their attention given everything that was going on proved to be a real challenge. Cabinet Office secured interest from Ross Clark (Scottish Government), Gary Haggarty (Welsh Government) and Richard Grace (Northern Ireland Government). My overall impression at that time was that interest in the capability was luke warm, despite the London based teams doing the heavy lifting to put the service in place.

Worth saying that from a user research perspective, both in terms of those potentially sending alerts, and the public receiving alerts, was conducted with participants from across the UK.

In relation to the responding agencies themselves, I presented to the Met Office, Natural Resources Wales (NRW) and the Scottish Environment Protection Agency (SEPA), the latter two knowing for some time that this capability was coming. Flood risk is a devolved responsibility therefore these Agencies have their own systems and autonomy.

It did feel at the time that the national alerting capability to some degree was seen more as a threat than an opportunity. Some time earlier, as the Service Owner for the Environment Agency’s flood warning service, I was requested by Natural Resources Wales to remove their warnings/data from GOV.UK and they only wanted this information on their own website. The Emergency Alerts service, effectively reverses this position given the objective of displaying any Emergency Alerts in force across the UK on GOV.UK.

One system, two routes to Alerting

There are two methods of using the Emergency Alert capability. Let’s call these 1) Through the front door & 2) Via the side door.

The front door is essentially having a login direct to the Emergency Alerts system. Role based access is employed and we elected to split the role for preparing the Alert and the Approver. In this example below during testing, Pete Herlihy (Product Manager for Notify) created the Alert and I approved it.

Setting up users is a fairly straightforward task and the process of writing a message, selecting the location and issuing is efficient. As with any digital service, our goal was for a user to be able to successfully sending an Emergency Alert, first time, unaided.

The side door is essentially using another system to generate the Emergency Alert. I was keen to develop this capability for two reasons: firstly, it enabled organisations such as the Environment Agency to use their own system preventing the need for two logins and saving valuable time during emergencies. Note that even if an Alert is generated from another system, it still requires approving via the GOV.UK service.

Secondly, it amplifies the benefits of using a national standard for the exchange of emergency alerting information known as the Common Alerting Protocol (CAP).

Shout out to Simon Nebesnuick and Ray Hanson whom I worked with to press for the adoption of CAP XML as the UK standard for emergency alerts. Any qualifying organisation adopting this standard could work with the GOV.UK Notify team to integrate their warning system with the Emergency Alerts Service.

At the Environment Agency we pressed on working on this integration. Shout out to Steve White, Ray Hanson, Claire Kemp, Chris Hill-Scott at GDS and Pete Davies at Fujitsu (plus many others!) to get this up and running. In the images below we are testing this in the development environment (2021).

Floods don’t follow administrative boundaries

The Side Door route allows for the importation of a‘target area’ into the Emergency Alerts Service. As we know floods don’t follow administrative boundaries, so these irregular shaped flood warning areas, which align to the nature of the flood risk, can be consumed and translated into an instruction to the mobile networks to broadcast an Alert based on this boundary.

When I left the national Emergency Alert Team, the Front Door route only allowed for the selection of Administrative boundaries not irregular shapes e.g. Local Authority or ward level or national, England, Wales, Scotland, Northern Ireland etc. We’d spoken with the GDS Notify team about the need to be able to draw a Targeted Area dynamically within the service, but I don’t believe that capability is currently available — happy to be corrected by any of the current team on that!

The hardest work

As a supporter of the Common Alerting Protocol for many years we’d already capitalised on the benefits of using this through working with Google. This put the Environment Agency in a prime position to use the Side Door for Emergency Alerts.

NRW & SEPA were aware of CAP but didn’t show interest in adopting this standard and therefore when the Emergency Alerts service became available, their only option was using the Front Door route*

*NB this was the position in 2021, my assumption is the status quo still exists.

Arguably, the hardest work is not necessarily building the capability, it’s actually working through how you want to use it, when you want to use, practicing and testing this so when the time comes, you’re ready.

Here’s some of the principles the Environment Agency was working to shortly before I left the organisation in April 2023.

I’m aware the Environment Agency brought in RAB Consultants to help them develop this work further. I haven’t seen any indication, at least publicly, that NRW or SEPA have worked through how they intend to use the Emergency Alert service.

In April 2023, SEPA tweeted that the Emergency Alerts system won’t cover flooding in Scotland at this stage and does not replace SEPA Floodline.

But what about the Met Office?

I’d say there was even more of a luke warm reception…possibly tepid at best from the Met Office when first introduced to the capability. Two reasons, firstly, having a strong commercial arm, adoption of open data and common alerting protocol has never really been on the agenda. If weather warnings were open data, others could re-use and integrate and innovate with that information. Secondly, Met Office has always steered away from direct alerting the Public. Their warnings are received by the media and responding organisations who then either relay these warnings or combine with their own intel and this appears within their direct warnings to the public.

However could the Met Office use the Emergency Alerts Capability? yes they could but via the Front Door Route they would only be able to select administrative boundaries and without the CAP integration for the Side Door route, the level of accuracy trying to manually match up a red weather warning to administration areas could potentially be low and therefore lead to over-warning.

So why was the new Emergency Alert Service not used in Storm Babet?

In summary, i’d say three things:

1. Low levels of interest and engagement from some of the administrations plus political will, both national and local, to make changes necessary to get the best out of the system through the adoption of common alerting protocols & standards.
2. A fear of losing autonomy/control and a feeling of a service being imposed on them — a national GOV.UK system at a time when each flood risk authority is wanting to diverge with its own systems for their citizens. This has led to delays with -
3. Devolved Administrations and their Agencies haven’t yet worked through how and when they intend to use it, overcoming the organisational inertia to make this happen.

If there had been sufficient threat (urgency, severity, certainty) in England, there would have been a much higher likelihood of the Emergency Alerts service being used operationally for the first time. I’d hope, once the flood waters have receded, there is a formal review of why the service wasn’t used in Scotland and this expedites progress removing the blockers to ensure this powerful alerting capability can be used when needed as it was designed to do, providing critical information to those who need it.

…

If you’d like to read more on the whole back story of how the UK Emergency Alert Service came about, you’ll find a three part blog starting here.