Quickly unfolding flash droughts present new challenge for weather prediction, March 3

By Michelle Klampe, OSU report – CORVALLIS, Ore. – Flash droughts, which can develop in as little as two weeks and intensify rapidly, are drawing new attention as researchers work to identify and test the reliability of indicators that could help predict these weather events, a study published today in Nature Climate Change indicates.
Understanding the factors, including a warming planet, that lead to a flash drought and improving prediction could have widespread impacts on people, said Oregon State University’s Philip Mote, a co-author of the paper. “We’re starting to understand that maybe we have some capacity to predict these flash drought events one to three weeks out,” said Mote, who is a professor in the College of Earth, Ocean, and Atmospheric Sciences and dean of the Graduate School at OSU. “That information could give reservoir managers an opportunity to adjust water level management strategies or allow farmers to adjust operations, for example.” The paper’s lead author is Angeline Pendergrass of the National Center for Atmospheric Research in Boulder, Colorado. Drought is considered one of the most complex and least understood of all weather and climate extremes. Droughts can last anywhere from a few weeks to decades and span a few kilometers to entire regions of the world. Their impacts often develop slowly and can linger long after the drought has ended. Flash droughts, as defined by the researchers, generally have three basic features: rapid onset and intensification; high rate of intensification; and a state of severity at the event’s end that would qualify it as a drought. Characteristics of flash droughts include a precipitation deficit, above-average temperatures and rapid decline of soil moisture. They may be triggered or exacerbated by heat waves. Flash droughts have always existed but scientists only began recognizing them in recent years and the new paper gives them a formal definition, Mote said. As the planet warms, droughts are expected to increase in duration and intensity, making understanding the causes and impacts more important than ever. In addition, more and more droughts have a temperature component; drought is exacerbated by warming temperatures, which can accelerate snowmelt or evaporation of moisture. In the past, drought prediction indices have largely been based on data collected monthly and then made available one or two months later. But that approach doesn’t effectively address monitoring for flash droughts, which can develop in as little as two weeks, Mote said. Since flash droughts have a shorter time to impact, forecasting becomes more important than for slower droughts, in which monitoring is the focus. In their paper, the researchers suggest that drought monitoring systems use shorter timescales and update more frequently. They also suggest using additional data, such as soil moisture and land-atmosphere interactions and their connections to large-scale meteorological conditions. “It’s partly a matter of looking differently at the information we have,” Mote said. “The cutting edge here is recognizing that the state of the land surface temperature and the state of the sea temperature – this kind of data can help improve our forecasts.” The researchers also noted a separate but related phenomenon: rapid-intensification snow drought, which occurs when snowpack declines suddenly. This type of drought is particularly concerning for regions like the Pacific Northwest that rely on snowpack for water supply and power generation. Understanding this type of snow drought is equally important moving forward, Mote said, but researchers agreed it is different enough from flash drought that it ought to be studied separately. Other future directions for the research include looking at past weather data using the new definition of flash droughts to see if actual events played out as they might have been predicted, Mote said.  “The goal would be to do research that helps us quickly improve our drought prediction models,” he said.