Colorado deploys latest flood forecasting technology

Colorado deploys latest flood forecasting technology

Published 20 March 2015

 

Colorado governor John Hickenlooper has raised concerns about House Bill 1129, a 5-year, $10 million proposal to implement a new technology which would assist with predicting the direction and intensity of wildfires and floods. Scientists have spent twenty years working on the proposed technology for Colorado, saying it is finally ready for implementation. The technology uses rainfall estimation, precipitation forecasting, and water modeling along with hundreds of thousands of atmospheric data points to predict the direction, speed, and intensity of floods, with up to 12-hour notice. As advancements in technology and computing allow for more accurate flood predictions, some variables remain a challenge for hydrologists.

Colorado governor John Hickenlooper has raised concerns about House Bill 1129, a 5-year, $10 million proposal to implement a new technology which would assist with predicting the direction and intensity of wildfires and floods. “In my experience, the ability to predict weather conditions and the types of information you need to predict fires accurately enough to be able to be really useful is very difficult,” Hickenlooper said. “One thing I would hate to do is commit $10 million to a fool’s errand.”
Techniques used to forecast floods have improved throughout the past few decades due to supercomputers, advanced radar systems, and satellites. In 1965 hydrologists had slide rules, rain gauges, and grave diggers. Pedro Restrepo, a 65-year old hydrologist responsible for the North Central River Forecast Center (NCRFC) in Chanhassen, Minnesota, relates to the tools used decades ago, while embracing modern technology. Restrepo began his career in hydrology in the 1970s, when the instruments being used were similar to those used in 1965. “I still have my slide rule,” Restrepo said. He compared the readily available data and computer models in his office to the challenges hydrologists faced decades ago.
On 19 March 1965, the U.S. Weather Bureau predicted there was a potential for flooding in Minnesota, but “the condition was not considered serious,” according to a government report released five years after the floods. By 30 March, the bureau predicted that “if snowmelt takes 3 days or more, Minnesota streams will not crest as high as 1952 …” which, according to the Mankato Free Press, meant the Minnesota River at Mankato should not top 24.6 feet.
It was standard practice in 1965 to use circumstances leading up to previous floods to estimate the severity of the current year’s flood, said Restrepo. Before the 1952 flood, a 1951 flood reached a level of 26.2 feet, but forecasters could not predict the 1965 flood that would ultimately crest at twenty-nine feet. By 7 April 1965, warm temperatures and heavy rains raised the bureau’s prediction for Mankato to reach 26.5 feet on 9 April. A day later, unprecedented flows of water led the bureau to push its prediction to thirty feet and said flooding would occur on 10 April. The actual crest in Mankato reached twenty-nine feet on 10 April.
Three days before the crest, the bureau’s forecast was too low by more than 2.5 feet and a day before the crest, it was too high by a foot.
Mankato’s last major spring flood threat in 2011 reached a crest of 25.5 feet on Saturday 26 March. Predictions from the (NCRFC) two days before the flood were accurate, but even with advancements in flood forecasting, the unpredictable weather left hydrologists worried earlier that week. On Monday, 21 March 2011, hydrologists predicted a twenty-seven-foot crest and increased it to twenty-eight feet on Tuesday when it began to rain. Both of those predictions suggested a Friday high-water mark, but by Wednesday, the crest forecast was reduced to 26.5 feet and the crest date was set for Saturday.
Boulder-based scientists at the National Center for Atmospheric Research have spent twenty years working on the proposed technology for Colorado, saying it is finally ready for implementation. The technology uses rainfall estimation, precipitation forecasting, and water modeling along with hundreds of thousands of atmospheric data points to predict the direction, speed, and intensity of floods, with up to 12-hour notice.
As advancements in technology and computing allow for more accurate flood predictions, some variables remain a challenge for hydrologists. For service coordination hydrologist Steve Buan with NOAA’s North Central River Forecast Center (NCRFC), the variables include: “the amount of snow on the ground in the drainage basins of all the rivers and tributaries, the amount of moisture in the snow (which can be different than the snow depth), the temperature of the air within the snow (if it is colder, it will slow the snow’s melt), the amount of moisture in the ground (dry ground will soak up snowmelt and spring rains), and the depth of the frost (water cannot soak into frozen earth).”
Noting that all of those variables pertain to existing conditions, Buan adds that flood forecasters also have to worry about the impact of future weather.