DATA MINING AND SPATIOTEMPORAL ANALYSIS OF EXTREME PRECIPITATION AND NORTHERN GREAT PLAINS DROUGHT

The Prairie Pothole Region in the Northern Great Plains (PPR-NGP) of the United States frequently fluctuates from drought to deluge. The weather events accompanying these disasters are characterized by large spatial and temporal variations from normal in the precipitation frequency and intensity. We performed data mining on available weather station observations to extract significant information and gain insight into PPR-NGP drought behavior.

Hydrometeorological indices calculated over various timescales were used to determine the controlling factors of changing drought conditions in the PPR-NGP. A series of analyses showed that large fluctuations in monthly drought conditions were controlled by changes in both the frequency and intensity of precipitation events. Extreme drought took a minimum of two to three months both to develop and dissipate from a semi-arid state. However, the biggest shifts in month-to-month conditions were observed when both precipitation intensity and frequency were well above or below normal over the same area; hence, the onset or end of an extreme drought took place more rapidly when the intensity and frequency changes were great and harmonized. The methods applied in this research could be adapted to a study of any other U.S. climate region.