Flagler’s Oddly Sunny Hurricane Matthew After-Action Report Draws Criticism From Its Own Emergency Staff

The Flagler County government administration today released its long-awaited after-action report on Hurricane Matthew.

It’s a surprising document in two tones: much of the 23-page report is a sunny, illustration-rich, back-patting recapitulation of the days of the storm, reading more like a public relations brochure than an after-action report. Its eight latter pages list 25 issues that emerged from the storm, with proposed solutions. That section more candidly outlines numerous problems with evacuation plans, the dissemination of information, lack of advance preparation, some lacking services at shelters, operational issues at the Emergency Operations Center, “gaps in financial and payroll procedures,” and other issues.

Yet the difference between the first and second part of the report creates a Jekyll-and-Hyde discordance that undermines the confidence in candor and self-criticism a report of this sort is intended to project—particularly to state and federal agencies that look to those reports when considering grants that may help address some of the problems pointed to.

The report as a whole  contrasts with—for example—the more sober, business-like after-action reports prepared by Palm Coast government and the Flagler County Sheriff’s Office, both of which provide much clearer roadmaps of the agencies’ actual “actions” before, during and after the storm than the county’s document does, connecting recommendations to a more logical context.

“We write reports differently from other people,” County Administrator Craig Coffey said, defending the approach and describing it as reflecting what was done well as much as what was done not so well. “We did a lot of things right but we could always do things better and that’s what we try to reflect.”

The report picks up what “action” there was close to the day the storm approached and passed by Flagler, on October 7, 2016, without providing any context for the several days that preceded the storm, when it was believed to be a potential direct hit on the county, and when Matthew was churning as a Category 4 and, potentially, a Category 5 hurricane. It was during that time—as the Palm Coast report notes—that many preparatory decisions were taken, when many potential deficiencies in the response were coming to light (not because anyone was at fault, but because such events inevitably bring to light unanticipated problems, which after-action reports are intended to analyze), and when the county was positioning its resources.

In an apparent reversal of roles, the county’s report was largely written by County Administrator Craig Coffey and his deputy, Sally Sherman, with input from departments, though such reports are usually written by emergency management staff and their department director (in this case, Steve Garten). The reversal underscores what had been evident at the time of the emergency: that Coffey took the reins of the emergency as, essentially, the emergency manager, with Garten in a role more akin to that of an executive officer.

Coffey says Garten wrote a portion of the report with and Michael Esposito, who’s in charge of special projects for the administration (he has no emergency management experience), and that a draft was turned in, but “we wanted to make it more critical of ourselves and broaden it,” Coffey said. “We wanted to expedite it too because it wasn’t being finished up as quickly.” So the final product was the result of “collaborative” work with input from numerous sides.

Emergency management staff was concerned about the final product, and it doesn’t appear as if it was given as much of an opportunity to participate as Coffey claims.

“I wish we would have had a chance to comment on this prior to it being minutes away from being released to the public,” wrote Laura Nelson, an emergency management mitigation specialist considerably more conversant with emergency operations than, say, Esposito. “I was under the impression we would get an opportunity to review it, particularly because the AAR will be utilized for our CRS annual report. However, it falls short and Adam and I will have to do an additional write-up.” (AAR is the acronym for the after-action report. CRS is the Community Rating System that helps communities potentially reduce their flood insurance premiums, but CRS points are earned based on creditable reports such as an after action report.)

In her notes on the document itself, Nelson wrote: “Document is missing some key information in my mind,” Nelson wrote, noting the absence of information on the days leading up to the storm, the number of people who applied for AID and the amounts distributed to “give an idea of just how impacted homeowners were with essentially a tropical storm,” how the flood warning and response plan was used, and what the report meant by “whole communities were flooded.” Nelson’s notes, by reflecting what was not included, point to the county’s skittishness regarding what could be perceived as bad news: “I think it would be a good idea to talk about how Fire Rescue had to stop responding when [the] winds got too high and how that actually impacted people who failed to follow evacuation orders,” Nelson wrote.

Coffey appeared not to have read the notes—he said he’d have to look at them before addressing them—then, when quoted Nelson’s words, said there was no reason to address Fire Rescue’s suspension of services when winds hit 45 miles per hour as that is protocol throughout the state and would not require addressing in an after action report. “Sometimes you get suggestions that are opinion, it doesn’t mean that was an issue to be writing about,”  Coffey said.

On a lesser point that nevertheless reflects the sloppiness that unnecessarily remained of the final product, Nelson also suggested pointed corrections, as when the report states in a self-congratulatory segment on social media, where the authors note that “Hurricane Matthew was the first hurricane to hit Florida since the advent and widespread use of social media.”

“This is not true,” Nelson corrected. “Matthew didn’t actually hit Florida and i would have considered Sandy to be one of the first widespread social media storms, which did affect Florida.” The correction was ignored.

The report includes numerous statements that seem out of context or less informative than boastful, as when it states that “Flagler County citizens were able to self-register for notifications using the CodeRED system.” The system allows residents to subscribe to weather and other alerts through their phones, email or devices. “Can we talk about how many messages were sent out via each? How many phone calls were made?” Nelson suggested. “For CRS We need to talk about the success of our mass notification system and this hardly talks about CodeRed at all.” The suggestion was ignored.

There are also factual mistakes that were left uncorrected, even though they were pointed out, as when the report claims the Flagler County Recovery Center “remained open until October 15.” In fact, it was open until November.

In an interview today, Garten, the emergency division manager, put it this way, when addressing the report: “I think it covers a lot of discrepancies that we have.” Asked whether the report was to his satisfaction, he said: “As far as being satisfied, I think it was as thoroughly done as it could be.” Emergency management staff “gave suggestions” to the administration, he said, who then put the report together. “Usually that’s how it’s done,” Garten said of emergency management writing the report. “In this case that’s what the county administration wanted to do, therefore following orders, we provided them with the recommended findings and they put it together.” The report may still be amended with recommendations and changes, he said, though the report the administration issued today was stamped “final,” even though it was also watermarked with the word “draft.”

Commissioner Donald O’Brien, who’s been pushing for the report to be released for a month in public and private meetings with the administration, was also struck by the tone of the report. “The beginning part of it was more descriptive and platitudes, but I have not gotten to the specifics,” he said. The part he hadn’t read was the last third of the report, which outlines recommendations. “Even without reading it I want to see more action plans rather than more description of things, for follow-up,” O’Brien said. “I’m probably not the only one thinking in that direction.”

O’Brien was under the impression that Garten and his staff had written the report, not the administrator and his deputy. “That’s not the way I would think it should work,” he said.

The report sheds little light on the command structure at the Emergency Operations Center, which was fully staffed for 144 consecutive hours, with 23 agencies manning desks and 80 to 120 people working a typical shift, plus 35 volunteers helping with meals, the call center, supplies and other needs. (It’s still not clear how many volunteer hours were logged.) At the emergency shelters, 1,814 residents got service, including, at one point—though the report does not say so—117 patients from Grand Oaks, the nursing home on Palm Coast Parkway, which initially refused to be evacuated even though it was in the evacuation zone.

In what seems to have been a legitimate reason for back-patting, the county mobilized a special medical support unit from Alachua County and successfully evacuated all the patients to a shelter—after pressuring Grand Oaks staff to let the operation proceed. But that detail was not in the report, which only vaguely refers to “some nursing homes” not evacuating. (Flagler County EMS, which presumably includes Palm Coast—the report does not specify—responded to 900 requests for service between October 5 and October 11, according to the report.)

The recommendations include developing a plan to ensure that traffic signals have ready generators during an emergency and planning for “comfort stations” in the county where people could recharge phones, email and do other operations prevented by a lack of electricity. Various recommendations address debris removal and improvements to shelter management. Several recommendations address Emergency Operations Center staffing and organization, including sheltering plans for staffers and immediate family members.

Coffey was impatient in an interview with questions about the report’s history and final form. (It is not, in fact, a “draft,” as suggested by the version sent out to media today, but a final report.) “You try and produce the best report you can and move on,”  Coffey said. Asked if this was the best report that could be produced, he said:  “It’s a good report.”

The full report is below.

Click to access Flagler-County-Hurricane-Matthew-After-Action-Report-6162017-Final-1.pdf