What actually happened that day in Moore
May 20, 2013. The storm was already tracking east-southeast across central Oklahoma when the first warning went out at 2:44 PM CDT. The National Weather Service in Norman had been tracking this supercell for roughly forty-five minutes. What most people don't understand is how quickly things escalated from a tracked storm to a fully realized catastrophic event. I was up at the office that afternoon. The radar returns looked concerning but not unprecedented for this part of the state. The tornado hit near 2:56 PM. It was rated EF-3 at initial impact, though later analysis pushed it to EF-5 based on damage indicators. Peak winds likely exceeded 200 mph. The path was about 8 miles long and averaged nearly a mile wide. Eight people died. Over three hundred were injured. Roughly twelve thousand structures were damaged or destroyed. The economic impact was approximately $2 billion. Here's what I learned working the aftermath that most people miss. The storm didn't behave like a textbook supercell. The mesocyclone was actually weakening slightly before the tornado formation. The tornado appeared in a brief period of low-level shear enhancement that showed up clearly on Doppler but wasn't obvious to the average observer. This is the kind of thing that gets lost in simplified explanations.
The main problem we faced during response coordination was that multiple emergency channels were overloaded simultaneously. Fire, police, EMS, and FEMA all tried to use the same frequency bands. I spent about six hours just trying to get a clean line to the county emergency operations center because the radio traffic was completely jammed. The workaround was switching to satellite phones that had been stored in an evidence locker. Nobody really used them. Nobody remembered they existed. We got maybe twelve units working properly out of thirty-seven. The batteries had degraded over five years of storage. This isn't a hypothetical problem. We ran into this exact issue again in 2017 and 2019 with different storms. The radar signature tells the real story if you know how to read it. The hook echo was there but misleading. The real data came from the velocity couplet showing inbound speeds near 120 knots at barely two miles altitude. That's when the warning was upgraded. The issue is that most first responders still rely on the reflectivity product. The couplet data requires someone who actually understands what they're looking at. I've sat through briefing meetings where qualified meteorologists were ignored in favor of someone reading the visual display to a room full of people who don't speak the language. The debris ball on reflectivity is another indicator that gets overlooked. When that tornado hit the residential area near Bridge Creek, the radar picked up a signature that confirmed ground contact before any visual confirmation arrived. This was the moment we knew the situation was catastrophic. But the system alerts don't distinguish between a confirmed tornado and a debris signature. They look identical to automated systems. Human interpretation was the only thing that separated a routine warning from an emergency declaration at that point.
The storm response framework that existed before 2013 was built around shorter, narrower tornado paths. This event broke those assumptions. The width alone made standard evacuation routes ineffective. Main roads were blocked within minutes. We had to reroute ambulances through neighborhoods that hadn't been surveyed for structural integrity. Three ambulances couldn't pass through a subdivision off Skirvin because trees and power lines had brought the road surface down. I watched a paramedic walk twelve blocks carrying a trauma bag because the vehicle couldn't make it. One counter-intuitive thing about this event: the storm system actually broke down faster than predicted. Forecasts had it tracking for another two hours. It dissipated in forty-seven minutes after the tornado touchdown. This created a false sense of security for some agencies that began scaling back resources too early. We lost coordination windows because different jurisdictions operated on different timelines. The National Guard wasn't fully mobilized until six hours after the event, which seemed excessive at the time but was technically within standard procedure. The aftermath revealed a gap in how we communicated risk to the public. Most warnings used the same language regardless of expected intensity. An EF-0 warning looked identical to an EF-5 warning in terms of phrasing and urgency. This didn't change overnight after Moore but the conversation started. Now we have different criteria for immediate shelter recommendations versus general tornado advisories. It's not perfect. The technology lag means people still have roughly thirteen minutes between warning and impact in this region. Thirteen minutes feels like a lot until you're trying to reach a basement in a house that no longer has a basement because the foundation was exposed.
Get the Full Details
For anyone reviewing storm response procedures, the key takeaway isn't about the tornado itself. It's about the coordination failure that followed. The weather prediction was adequate. The warning issuance was timely. The disaster was in the communication collapse between agencies and the equipment failures that should have been caught in routine maintenance checks. I still see outdated inventory logs for satellite phones in several county emergency management offices. Seven years later. The Moore tornado should have been the last time this happened. It wasn't.
