Support Provided ByLearn More
Planet EarthPlanet Earth

New System Predicts Flash Droughts Up to Four Months Ahead of Time

ByGreta FriarNOVA NextNOVA Next

Receive emails about upcoming NOVA programs and related content, as well as featured reporting about current events through a science lens.

The 2012 U.S. flash drought that cost the country $30 billion in economic losses could have been predicted up to four months in advance.

Scientists at the National Center for Atmospheric Research (NCAR) have analyzed soil moisture and snowpack data in the months leading up to the drought and have discovered that they could have been used to make a much earlier

Support Provided ByLearn More
drought prediction . If the model can be applied to similar droughts in the future, it could have huge impacts on crop yields, food prices, water management, and perhaps even taxpayers’ wallets.

After the 2012 U.S. drought, taxpayers footed the bill for $17.3 billion in federally subsidized crop insurance payouts to farmers. Prices for produce and meat shot up in response to low crop yields, with effects lasting several years and—because the U.S. is a major food exporter—affecting the entire globe .

If droughts could be predicted before planting season, farmers could change their irrigation and planting schedules.

Current drought forecasting systems cannot predict flash droughts like the 2012 one because these droughts are not preceded by typical indicators, such as large shifts in weather systems that lead to less precipitation. The new model, developed at NCAR, looks to the ground as well as the sky for clues to an upcoming drought.

“We need to understand the land surface conditions in addition to understanding the overlying climate conditions,” said NCAR scientist James Done, co-author of the study.

Case in point, soil moisture affects the exchange of water and heat between the ground and atmosphere, which in turn affects weather patterns and precipitation. Due to that connection, an earlier study found that low soil moisture could be used to predict flash droughts one to two months in advance. NCAR scientist Debasish PaiMazumder, along with Done, wondered if that model could be improved upon.

PaiMazumder, who moved to Boulder, Colorado, in February 2012, months before the drought hit that region hard, recalled being surprised by the mild winter. He wondered if the lack of snow in February could have been related to the drought conditions in the summer. He and Done analyzed decades of data on snowmelt, soil moisture, and drought. They found that low snowmelt in the spring correlated with dry soil later on and could be used to extend the previous drought prediction model to four months lead time.

The research has not been incorporated into active weather forecasting systems yet. The process of turning research results into prediction tools can take around six months, Done said, based off previous experience.

Earlier drought prediction forecasts “will give more time for farmers, ranchers, and water resource managers to prepare for this type of drought,” PaiMazumder said. “Lead time could be important for mitigation strategies.”

During the 2012 drought, 88% of U.S. corn crops were affected. Yield decreased by 50% in the cornbelt, causing corn prices to rise by around 20%, said Gary Schnitkey, a professor at the University of Illinois who researches farm and risk management. If droughts could be predicted before planting season, farmers could change their irrigation and planting schedules, in accordance with drought mitigation strategies . They might plant more drought resistant crops instead of corn, Schnitkey said.

In spite of the new research, Schnitkey was skeptical of the prospect of an accurate long-range drought forecast system, at least in the unpredictable climate of the U.S. cornbelt. “There really isn’t a model that most people believe or that has been proven to be accurate” for predicting weather more than a few weeks in advance, he said. A “false positive” in which farmers cut back on crops or culled cattle in preparation for a drought that never came would be a serious problem, Schnitkey said.

The NCAR team is optimistic that their research can be used to improve existing drought prediction forecasts, but agrees that advanced weather forecasting is challenging and that lots of unknown factors remain. PaiMazumder and Done are already searching for more drought predictors that could be added to the model, as well as for indications of how “flash drought is changing in changing climate,” PaiMazumder said.

The frequency of flash droughts is expected to increase in the coming years, so NCAR scientists are working with a sense of urgency to build better models. “Watch this space,” Done said.