The Marshall fire swept into Louisville and Superior, killing two people, forcing mass evacuations and causing $2 billion in damage not just because of the intense flames, but also because it raged amid an all-day hurricane, the nation’s top weather analysts said Tuesday.
While Boulder and Jefferson counties see a handful of wild downslope winds each year, the storms almost always gust their dangerous blasts and then offer relief by falling into periodic lulls, according to a new study by Boulder researchers from NOAA and the National Weather Service. But the Marshall fire on Dec. 30, 2021, blasted well over 100 mph and stayed at hurricane force for 11 straight hours, they said.
The scientists call it “nothing short of miraculous” that only two people died as winds pushed a wall of flame dozens of miles east into packed suburbs, in a study published in December in Weather and Forecasting, a journal of the American Meteorological Society.
A western Boulder County sensor at noon that day, about a half-hour after the fire ignited in two places, detected gusts at 57 meters per second, or 128 miles per hour. Peak gusts near the heart of the fire hit 115 mph.
The study does not claim better research could have made a difference that long day. The investigation showed the fire sparking at two spots from separate causes, and the hurricane winds overwhelming even the best preparation. Evacuation communication was imperfect, but people managed to get out of the way.
Illustrating the unicorn nature of the day, the researchers’ slideshow included snapshots of residents holding fist-size embers carried miles east of burned homes and landing in Boulder County yards and streets.
But the NOAA-led group said fire reviews have led to an easing of rules for when to call an official red flag warning that can broaden emergency alerts and media coverage ahead of a possible spark. While a high wind warning was issued before the Marshall fire, the higher-profile red flag threshold at the time was not reached because relative humidity stayed well above the 15% trigger levels, researchers said.
A new warning protocol allows weather analysts to put up red flags based on unusually high winds, even if humidity is not a factor, they said.
The researchers also said new modeling and geographic targeting tools can let weather analysts draw a precise “polygon” around the cellphone towers that can then issue emergency alerts directly to all cellphones within the temporary boundary, just as in an Amber Alert.
Marshall was “an unbelievable event,” said NOAA Global Systems Laboratory researcher Eric James. “Even without the fire, the wind would have been a major story.”
“Once the fire started, we knew it was going to be awful,” another researcher said. The junction at Colorado highways 93 and 72 south of Boulder is the windiest site of all NOAA’s sensors.
Downslope wind waves “crash together” there, they said. “Then, we got two fires.”
The researchers studied whether climate change, which has raised average Colorado temperatures 2.3 degrees Fahrenheit since 1980, might have factored into the Marshall fire.
Records in fact indicate there are slightly fewer dangerous high-wind events coming down from the Continental Divide to Front Range counties in recent years, they said. Climate change in many cases is said to make off-the-charts weather events more likely, from drought to torrential rain to deadly urban heat waves.
There are actually two fewer Front Range wind event days a year recently, though useful records don’t go back as far as researchers would like, said Paul Schlatter of the National Weather Service Boulder office.
More relevant for the Marshall fire was “not having any snow cover in the previous month,” researchers said. That December had been extremely dry for many Front Range counties. Moreover, the drought desiccated grasses and shrubs that had shot up during a relatively west spring of 2021.
Continual improvement in forecasting power should make future warnings more precise, for all kinds of storms, the researchers said. One noted that forecasts hours before Marshall showed typical downslope threats centered farther north, in Larimer County.
The biggest danger turned out to be “only half a county farther south,” NOAA’s Stan Benjamin said. “But it’s a big county.”