IGEM Report 1: 2017-18 - THE CYCLONE DEBBIE REVIEW

The Cyclone Debbie Review

In Debbie, more than 5.9 million messages were sent across Queensland during the event – about one quarter of all EA messages sent nationally since the system was introduced in December 2009. 109 About 3.6 million of these were for advice on school closures. The majority of campaigns used both the wide-coverage SMS-to-handset-location and voice-to- landline options. Only three campaigns used the more precise SMS-to-registered- service-addresses. This was to warn of the slow-onset flooding of the Fitzroy River. 110 The EA performed within expectations for getting messages to people for the majority of campaigns. The average successful delivery rates of the QFES campaigns were 88 per cent for SMS to handset location, 60 per cent for SMS registered service address based and 63 per cent for voice. 111 As a rule of thumb, i the average successful delivery rates are greater than 90, 70 and 60 per cent, respectively. However, for a number of campaigns the performance was less than normal expectations:

EA was also used to advise about school closures. If decision makers were aware of how long this would take – more than four hours from an email that showed the decision had been taken to when the last EA campaign ended – they may have decided to warn in other ways.

Themes

School closure

The biggest Emergency Alert campaign undertaken was to communicate the decision to close all schools south of Agnes Water and west to Nanango. While it was identified by most councils and agencies in the southeast as an effective strategy to reduce the volume and risks of people on the roads, there were problems identified in the implementation of the campaign in regards to timing, scope and clarity of the messaging. The decision to close was taken early. At 7.32am the Department of Education and Training tweeted that “all state schools in Metropolitan, North Coast and South East Regions are closed.” An email to all South East Queensland public servants at 8.35am on Thursday 30 March told readers that “all schools from Agnes Water to the border of New South Wales, and west to Nanango are closed today.” But the first EA campaign did not start until 9.39am. Given the length of time to issue 11 campaigns to more than 3.6 million people, some EAs did not commence until 10.50am, and finalised at 12.50pm. As timing was a critical factor to ensure children did not end up at schools or in transit, the use of other more time efficient methods would have enhanced the process. Many people were aware of the closure hours before via radio, Facebook or websites. This emphasises that EA is just one of a suite of potential messaging tools, and circumstances should influence which is the most appropriate. Many councils reported high volumes of calls from people seeking clarity on the decision having heard the information via radio or Facebook. As one council identified: “council was not consulted and found out at the same time as the community. This created confusion as people began ringing the council call centre who at that stage didn’t know what was happening. They received inconsistent advice as no formal information to use it made the state and local look at odds.” Most councils also identified that the timing of the message as critical, identifying that children were already at school or in transit, and in particular noting that rural school buses often commence by 7.00am. The Queensland Evacuation Guidelines for Disaster Management Groups strongly recommends that the community is informed of school closures as early as possible and preferably before the school day starts, given the impact for carers. In particular the Guidelines note that an early decision will reduce traffic on the roads and the need for carers to return and collect children from school. Better coordination of the message, scope, timing and best method for informing the public would have greatly have enhanced the implementation of the decision. Given that the decision to close the schools was made at a state level and not at the local level the full utilisation of the CCN may have also strengthened the communication around the decision.

• Eleven campaigns with the voice option reached less than 50 per cent of landlines. • Less than 60 per cent of the school closure campaigns were successful due to the size and concurrency of the campaigns exceeding either the capacity of the system or campaign limits.

While the message may have got through in many cases, its timeliness is imperative when warning the community ahead of an impending event. Any delay in delivery can pose additional risks to community members. We heard mainly positive feedback about the timeliness of campaigns. Our community survey showed 65 to 85 per cent of respondents who had received a message stated it had arrived at the right time. However, a few local government stakeholders thought some campaigns took longer than expected, as did some respondents. One local government stated the centralised process for issuing campaigns led to a backlog and delay of their campaign.

Notes i. Based on the averages of campaigns since the introduction of EA.

93

6 THEMES

Emergency Alert process diagram • School closures