News & Announcements

Climate typically is computed from meteorological data sampled over 30 years. Climate models do not model the climate directly; they compute solutions to equations of evolution for the Earth system’s instantaneous state and write daily or monthly summaries of that state to history files, which are then postprocessed to compute climatologies. The I/O-intensive nature of this process is arguably the most significant barrier to creating an exascale climate model. A natural question arises: Can we model the climate directly?

Jay Larson, a computational scientist in Argonne’s Mathematics and Computer Science Division, has developed a data analysis and information theoretic framework to analyze climate variability and applied the new framework to the Central England Temperature (CET) record, which commenced in 1659 and is the longest weather station-based observational record available.

His approach, based on estimating probability density functions (PDFs) for each 30-year sampling window for the record, provides deep insight into the time evolution of CET monthly and daily average and daily extreme temperature climatologies. The time-dependent PDFs are consistent with and add new detail to the CET’s known warming trend. These PDFs also show oscillatory behavior that may be connected to known interdecadal and century-scale oscillations present in the CET monthly average time series.

Information theoretic techniques have been applied to compute how “unusual” any 30-year climate sampling interval is with respect to the full CET record and that of the pre-industrial era (before 1870). The climatology of CET average, maximum, and minimum temperatures in recent decades differs distinctly from that of previous observed times; these differences are more pronounced than for any other 30-year period in the CET. 

Furthermore, climatologies computed from recent decades are distinctly different from any other 30-year period from earlier parts of the CET. Taken together, these analyses identify much of the 19th century as a period of relative climatic stability (i.e., slow changes with respect to time), and identify the past 30 years as a period of rapid climate change.