Guest Post: The Unnecessary Tragedy of Hurricane Helene

As of October 23 2024, 224 people had died from Hurricane Helene, and 24 were still missing. And, although some amount of tragedy is unavoidable in any natural disaster as ferocious as this storm, flawed decision-making led to unnecessary fatalities.

What’s the problem, and how can we fix it?

It turns out that the “cry wolf effect in evacuation” is part of a larger and systematic flaw in decision making that is, unfortunately, increasingly common: the inability to adjust behavior in the face of changing circumstances. As published in Physica A:

…false alarms are becoming more and more common. Compliance to an evacuation order made by an authority in case of emergency can play a key role in the outcome of an emergency. In case an evacuee experiences repeated emergency scenarios which may be a false alarm (e.g., an evacuation drill, a false bomb threat, etc.) or an actual threat, the Aesop’s cry wolf effect (repeated false alarms decrease order compliance) can severely affect his/her likelihood to evacuate.

Learning from the past inhibits adaptability to new circumstances

Was the cry wolf effect in play during Helene evacuations? Apparently. Early on I saw mentions on social media of people who said they’d been fooled by too many prior flood warnings. Checking facts, I found a local story quoting a local meteorologist:

According to WRAL meteorologist Chris Michaels, flash flood emergencies are extremely rare. “It’s pretty rare. In 2023, less than 1% of all flash flood alerts in the country were flash flood emergencies,” Michaels said.

And therefore, people dismissed the evacuation order as just another nothing-burger alert.

This is a decision pattern that is popping up everywhere, with devastating effect as with Helene and so many more circumstances lately. Why? Think about it: we tend to think about decisions in terms of how my actions affect the outcomes I’m seeking. We make an implicit assumption that the world is stable, so we can mostly ignore external factors like weather or climate. This has worked for us because for most of our lives (indeed for most of human history) we have lived in a world where externals didn’t change much or only changed predictably or seasonally.

We have trained our intuitions so that we make good decisions in this world. When the externals change dramatically, our intuitions fail us leading to unintended and sometimes tragic consequences.

False alarms inoculate us against the real thing

People in Helene’s path had seen many flood warning false alarms, most of which involved rain after a period of relative dryness. But Helene violated two flood assumptions: it was a fundamentally new situation. First, according to phys.org, Helene’s rain came on already very wet ground. Second, East Tennessee State University explains that there was a huge volume of rain in a short time period:

Rain fell quickly at the upper reaches of three major watersheds: the Nolichucky, French Broad and Pigeon rivers. The result was a torrent of water and debris racing unusually fast down narrow channels and valleys… And the life-threatening flash flooding in Northeast Tennessee extended beyond the traditional “100-year” flood boundaries that weather officials use to depict flood vulnerability.

There’s a general principal here that goes well beyond disasters like Hurricane Helene. When externals are no longer stable—when climate change makes hurricanes more dangerous, a drought cuts Panama Canal traffic by 29%, or supply chain risk increases due to conflict, deglobalization, or environmental concerns—how can we reeducate our intuitions?

The bottom line: we live in a world of increasing instability, and so we have to reason more carefully: “what’s always worked” thinking can kill us.

Far from the dangers you might read about in the news: AI and DI can go a long way to help

Decision Intelligence, as I wrote about in the DI Handbook with L. Y. Pratt, gives a step-by-step approach to overcoming this kind of reasoning limitation.

First, we can draw a decision diagram where we explicitly name the important externals and identify the causal chains that link them to the outcomes we want. Sometimes a simple picture (Causal Decision Diagram in the book) can tell us that something is worth a quick check.

You can also use an LLM to help. I asked chatGPT, “Has hurricane damage, Panama Canal throughput, or supply chain risk changed in the last 10 years?” and in seconds got an answer consistent with the sources cited in the first sentence of this paragraph. LLMs make mistakes, but this kind of response to a quick check should be enough to warrant more digging and thinking.

Second, when externals key to our lives or business operations are changing, decision intelligence simulation shows us risks, mitigations, and optimizations under many different sets of conditions. We can ask how much Panama slow downs will decrease our profits or how delayed parts shipments will impact fleet maintenance. We can find the crossover point where canal costs or delays or supply chain instability make using a more expensive locally made item is more cost-effective than a cheaper foreign one. Or we can identify trigger points—when we see a specific external cross a threshold—that tell us it’s time to consider changing an action.

Take a few minutes to identify the key externals that can change your business success, your quality of life, or your life. Ask an LLM how volatile they are, ask the LLM for its sources, and at least spot check that LLM. Draw the CDD showing your actions, externals and outcomes: even a simple one can go a long way.

Do you need to re-train your decision intuition? It could be a matter of life or death.