Before Hurricanes Matthew, Irma and Dorian approached Flagler County in three of the past four years, the county scrambled to take protective measures on massive scales, as did most counties in the state, as did the Panhandle when Hurricane Michael struck the one year we didn’t have one here. The preceding warnings from local and state officials, including two governors, were dire, intended to scare people into complying with precautions, sheltering in place in most counties, evacuations in others.
The emergencies were disruptive, closing schools and businesses and altering lives, not just before the hurricanes neared, but even after they’d long passed, with power cuts lasting a week after Matthew and Irma. But even Hurricane Michael, which was extremely powerful but found the path least inhabited in the Panhandle, the expected disasters did not take place. Flagler had at various points feared a direct hit from Matthew and Dorian. A nearly imperceptible shift in Matthew’s path and a slightly more pronounced for Dorian one could have obliterated Flagler Beach. But both followed nearly the same path, wobbling just enough offshore to spare Flagler Beach and Palm Coast anything like a direct hit, even if both left enough damage as it was.
Yet for days, various models had predicted a direct hit, either here or not far from here. We had the European model, the American model, and dozens of models on top of those, creating the famous spaghetti effect, with outliers and everything in between. The American model was a synthesis of many others, producing NOAA’s famous cone of uncertainty–a modern equivalent to the oracle at Delphi.
Were the models wrong? At various points in the hurricanes’ paths, some of them proved wildly wrong, some of them not nearly so, some of them ultimately proving quite accurate as the forecast window narrowed, and one or two, such as the European model, were uncannily accurate even at several days’ remove.
Does that mean models are useless? Of course not. The models did their job. They’re never presented as dogmatic truths from some meteorological pulpit. They’re presented as probable paths, with that famous cone of uncertainty extending far and wide to account for any hurricane’s temperamental character and serve primarily as warning systems, not predictors of certainties. Officials use them to make sure we take precautions and to minimize loss of life.
The same principle applies to modeling a virus’ potential in a pandemic, though unlike hurricane forecasters, who’ve been dealing with the same set of data and variables for decades, scientists have less reliable data to go on with the novel coronavirus, and minimal historical data to gauge the effects of social distancing.
So it’s been strange to see Gov. Ron DeSantis ridicule Covid-19 models since late April, and more especially last Thursday, when he announced his otherwise cautious plan to reopen the state. His overblown rhetoric about the models, as if he were looking for an enemy or a scapegoat, did not match his commendable caution. Nor did it match his own reliance on those same models weeks earlier, when he used them to impress on Floridians the urgency of closing things down and sheltering–just as he would on the approach of what could have been the most dangerous hurricane in the state’s history.
As it is, the coronavirus emergency hasn’t been less than that. Hurricane Covid has so far caused more deaths in Florida alone–1,539 at this writing–than all hurricanes combined everywhere in the United States in at least the last 20 years, with the exception of Hurricane Katrina, which took about as many lives. That’s the official count, though the actual count is certain already to be much higher. By the time the tally is over, deaths are expected to be double or triple the current tally. That’s just for Florida.
So the governor’s triumphal show of charts and chest-thumping last week and his ridicule of what he considers unjustified fear is not only at odds with the actual death toll, but also with the alternative reality he’s re-imagining. Is he regretting following the models’ lead? Is he suggesting that Florida should not have shut down after Covid made landfall in South Florida, and accelerated through visitors from the New York region? The models he’s been picking on since last week in any case were the outliers (“The governor used some really extreme examples that I would say, wow, that wasn’t observed anywhere to that degree,” Daryl Tol, president and CEO of AdventHealth Central Florida division, said in a news conference the day after the governor’s announcement on opening)–a clever bit of intellectual dishonesty befitting DeSantis’s lawyerly skills but not the public he serves. He never acknowledged that the good models change daily, with new data. Nor did he acknowledge that he was citing models dating back to before social distancing rules were imposed–the very same rules that were intended to reduce the models’ projections, and that, in fact, did the job.
“The smartest models are updated every day or every few days,” Tol said. “The fact they shift is because they’re smart, not because they’re wrong. If they didn’t shift, there’d be a major flaw in the model. The great thing is human actions and leadership actions shift models because human actions change reality and reality changes the models…. I think you’d want us to be prepared. We’re the kind of place that isn’t going to say, well, the hurricane will turn away. We’re the kind of place that’s going to say: all right, if it hits us, will we be ready? And that’s how we’ve used models.”
To discredit models is as absurd as if DeSantis were to discredit public health agency recommendations or their army of contact tracers. Why he felt the need to do so suggests a couple of impulses are at work: he wants to ensure a safe reopening of society. That’s not in doubt, and so far he’s taking a more responsible approach than a few other, more hurried states. But his attacks on models also panders to the appealing but false logic that because the models’ predictions didn’t come true, the models were wrong. He’s been quieter this week since models have again shifted. One of the more reliable models raised the tally of expected cases and deaths in Florida to nearly 4,000 now that the state is moving, however carefully, toward reopening. The national tally is in the 130,000 range, nearly double today’s total of 74,000.
I doubt DeSantis will be ridiculing the next models come hurricane season, or the next Covid wave, should there be one, though he’s doing a good job of undermining their validity, and along the way undermining the mission of public health and emergency management workers. Unless of course he thinks that, whatever the models did or did not say, 1,539 dead Floridians is something to brag about.