You are here
The 2017 spring and summer drought over Montana, North Dakota and South Dakota sparked wildfires and reduced agricultural production, resulting in economic losses that exceeded one billion dollars. Neither the drought’s swift onset nor its severity were forecast. This report explores the causes, predictability and historical context of the 2017 drought over a Northern Great Plains region.
The rapid onset of drought in spring and summer of 2017 was mainly due to the failure of rains. Rains failed as a result of persistent high pressure that deflected moisture-bearing storms away from the region during what is normally the wettest time of the year. Periods of hot temperatures, high winds and low cloud cover also contributed in the drought’s rapid onset.
Two main characteristics of the 2017 drought were record low precipitation and near-record rapidity of the soil moisture decline. Observed May-July precipitation over eastern Montana was the lowest on record and precipitation averaged over Montana, North Dakota and South Dakota was the third lowest on record dating back to at least 1895. Between mid-May and early June, eastern Montana estimates of soil moisture declined from the 80th to the 17th percentile, the third largest such decline for any three-week period over that area since at least 1916.
The longevity of the 2017 drought paled in comparison to many prior droughts over the Northern Great Plains. The 2017 drought ended just three months after it began, and was thus considerably shorter than other standout droughts that include the ‘Dust Bowl’ of the 1930s and epochs during the early 1920s, 1950s, early 1960s, late 1970s, and late 1980s.
Climate model simulations reveal that anthropogenic forcing made the occurrence of the observed 2017 drought intensity up to 20% more likely. The increase in drought likelihood during July from a past to the current climate in model simulations is due to long-term reductions in soil moisture, also referred to as aridification. Aridification is forced by increases in evapotranspiration associated with rising temperatures.
Below average May-July 2017 precipitation was not predicted in advance of the season. State-of-the-art seasonal prediction systems initialized in April 2017 forecast that the observed low Northern Great Plains precipitation was an unlikely occurrence.
Cumulative precipitation deficits were only predictable through sequences of up to three day forecasts. Sequences of longer than five day forecasts provided no indication that the seasonal evolution of precipitation would be different from average.