The Problem of the Last Mile
Warnings are pretty good nowadays, people trusting and acting on them is not.
You are probably aware of the flash flooding catastrophe in Texas over the weekend, where way too many people were killed as floodwaters rose over 20 feet in just over an hour in some places.
Based on all reports, the forecasts of this event days earlier were less severe than the actual outcome, but the watches and warnings were issued well in advance with the right amount of urgency needed to take action in a life-threatening situation.
A tragedy like this one is preventable with the right action plans built on the trust of the warning process. That, unfortunately, didn’t seem to be in place for this event. As we look into the future, I’m alarmed that many things will likely continue to combine to create a toxic mix of conditions that will likely lead to more tragedies in the future: more frequent extreme weather events, less trust in institutions and science, critical understaffing and under-resourcing of the NWS, and a fragmented media system where truth is relative and facts are not believed.
Already, we’ve seen a massive rise in ‘weather control’ conspiracies to explain away the increasing frequency of severe weather events even when climate science is right there already doing the work of explaining the why behind many major events. These conspiracies have the effect of weakening the foundations of trust in an enterprise that is working extremely well; good warning information was built on strong science, and improving forecast data.
The simple truth is that our weather models have undergone significant improvements over the last decade. However, they still often miss heavily mesoscale-driven events until the final hours leading up to them. Similar to the flooding with Helene, there were hints of an excessive rainfall event in the days leading up to July 3-4, but it was in the final hours that the HRRR started to indicate something historically significant was possible in an area already well-known to be susceptible to flash flooding. The flood watch issued at 1:30 the afternoon before the event didn’t mention 10-15 inches of rain (or more), but the watch was still issued hours in advance with a pretty excessive 5 to 7 inches of rain mentioned.
One of the most challenging aspects I see forecasters struggle with is the problem of providing the last mile of warning information. This is the point when a watch or warning has been issued, but those in the watch area may or may not receive that critical information and/or take it seriously.
Regardless of what the watch text says, almost everyone under it will not read the actual text, and an alarming number of people will not even know a watch is in effect. So, regardless of what the text said, the number of people already weeded out by the last mile of warning information breakdown is already catastrophically high.
When the flash flood warning was issued 12 hours later, the text itself reads as if immediate action is required for those in flood-prone areas, because it was.
This initial warning was delivered in plenty of time for almost everyone. If someone had gotten the EAS alert on their phone, they had plenty of time to act to get out of harms way. However, it should be noted that Texas is notoriously bad about using this system for a variety of non-emergency uses which could prompt people to turn them off and limit their effectiveness.
Additionally, another problem I want to highlight that probably had some impact is the rampant misuse of the weather info space by bad actors who proclaim high-end events are coming with every system, even when those events seem unlikely. Oftentimes this is info delivered on social media for the purpose of gaining attention, a valuable commodity — but it is also grossly irresponsible to do that over and over again because the fatigue that can build up in individuals who follow these notorious accounts is very real.
All of this highlights the core problem: the last mile of warning delivery converting into public action is the problem the weather enterprise will continually grapple with.
The Flash Flood Emergency was delivered as the massive flood wave was moving downstream. By all accounts, many areas saw river rises of over 20 feet in an hour or slightly more in the middle of the night, which is catastrophic.
The reality of the human loss from this event is grim. But what isn’t in question is that the weather service, forecasters, and the weather enterprise as a whole did their part in forecasting and publishing critical information.
This Has to Be A Wake-Up Call
This is an event that raises critical questions for our country in an era of rising extreme weather risks: What measures can we take to ensure people receive accurate information? What can the weather service do to ensure the right messages are being delivered? What can we do to ensure the weather service is staffed up and the science/research is improving? What public infrastructure can exist to ensure safety in the face of these dangers?
It isn’t without precedent that public accommodations in dangerous areas must meet safety standards to operate. You don’t board an elevator that hasn’t been inspected, or stay at a hotel without fire evacuation plans. It seems evident that weather readiness should probably be on that list as well.
Misinformation about the weather is rampant, bad actors are exploiting fear for attention, and outdated myths that are clearly incorrect are still widely believed. Oftentimes, there’s an air of mysticism around weather events, as if science doesn’t already have explanations for them. Education has to be a critical focus for all involved in the weather enterprise, at all corners, especially those with large audiences.
The last mile of weather information delivery will continue to be the hardest question facing the warning system. Still, an essential aspect of solving this will be educating people on how the weather actually works, taking the mystery out of our most extreme storms, so that people can have a framework of understanding to act within.
What happened in Texas can’t be allowed to happen again, and we have the tools to prevent tragedy in a majority of weather events. And as this event has highlighted once again, the warnings are oftentimes not the problem; it’s what happens after they are issued that we must figure out..