Forecast models did not accurately predict the timing, magnitude and exact location of heavy rainfall that caused massive flooding along the Front Range last September, according to a self-assessment released Tuesday by the National Weather Service.
"While National Center forecasts indicated the potential for very heavy rainfall and at least a slight risk for rainfall to exceed flash flood guidance, significant variability in magnitude, timing and location of the heaviest rainfall made it difficult for the (weather forecast offices) to anticipate the rapid evolution of the flash and river flooding," the report's authors wrote.
The September flood shattered weather records, claimed 10 lives in the state and caused an estimated $2 billion to $3 billion in damage.
The assessment was compiled using interviews and product evaluations to gauge the organization's performance before and during the weather event.
"I was very happy with the office's overall performance," said Nezette Rydell, meteorologist in charge at the National Weather Service in Boulder. "I personally believe lives were saved because people had warning information in their hands early enough to make a difference."
She said the extreme nature of the weather event made it difficult to forecast, with models showing multiple possibilities.
"It wasn't a given that we would have this much rain," Rydell said. "We didn't miss it. We just didn't quite have the magnitude."
As early as five days in advance, the National Weather Service's Weather Prediction Center began issuing 48-to-72-hour forecasts that included the potential for 2 to 4 inches of rainfall for portions of the Front Range.
'A lot of moisture in that washcloth'
Tom Hamill, a meteorologist with the National Oceanic and Atmospheric Administration's Earth System Research Lab in Boulder, evaluated the forecasting models for the report.
He said it was "a little bit surprising" that the models weren't more accurate or consistent given the high humidity preceding the torrential downpours.
"There was a lot of moisture in that washcloth to wring out," he said.
But, he noted, "forecasting heavy precipitation is an extremely complicated problem. We know we need better forecast models, and we need a program of research and development to do that."
Despite some obstacles and areas of improvements, most parties interviewed for the assessment were "satisfied" with the level of service provided by the National Weather Service during the flood.
"Although forecasting the specific timing and location of heavy rainfall and flooding was extremely challenging, the Boulder emergency managers said that the information provided by the National Weather Service helped them save hundreds of lives during the first night of the flood," said assessment contributor Rebecca Morss, a scientist with the Boulder-based National Center for Atmospheric Research.
Weather service helped 'immensely'
Mike Chard, director of the Boulder Office of Emergency Management, said that, from his perspective, the National Weather Service did a good job providing some advance warning and keeping his office up-to-date on rapidly changing conditions.
After the flooding started the night of Sept. 11, he said, his office was sharing information every few minutes with the National Weather Service's severe weather desk.
"It helped us out immensely as we were going through the event," he said. "We were able to compensate for any inaccuracies in the modeling by having that human interaction."
Other report findings include:
Limited coordination between the NWS Weather Prediction Center and weather forecast offices and river forecast centers regarding rainfall and flash flooding potentials. The report recommended replicating the "successful" winter-weather coordination model for future hydrologic events such as floods.
The NWS does not have a formal policy for what it calls decision support services, or information provided to emergency managers for their planning and real-time decision-making.
NOAA should work to improve its quantitative precipitation forecasts to better predict extreme precipitation events, and the amount of precipitation that could fall.
Slow Internet connections made it difficult for the NWS to stay up-to-date and provide current information to relevant parties.
The NWS should update outdated training modules and add more in-residence training for weather professionals.
The findings, meteorologist Rydell said, "will help guide the agency on where to invest their resources in the coming years."
Contact Camera Staff Writer Amy Bounds at 303-473-1341 or email@example.com.