By applying advanced data mining techniques to observed and model-simulated climate data, a researcher at Oak Ridge National Laboratory envisions the creation of a new set of tools that can provide valuable insights into climate science. Ultimately, Karsten Steinhaeuser believes this work will lead to improvements in predictive modeling and reduce uncertainty. "We are effectively using data mining techniques to ?learn' the not-so-well-understood physical relationships in the global climate system," said Steinhaeuser, who is earning a doctorate at the University of Notre Dame. While traditional approaches have focused on descriptive analysis, Steinhaeuser noted that this "unified framework" approach uses complex networks to discover relationships between ocean climate indicators and land climate from observed data.
Menu