CMG-Bayesian Hierarchical Models
for High-Resolution Climate Reconstructions

CMG Home

Project Description

 - Motivation and Goals
 - Project Outline

Collaborators

 - Project PIs and
    Collaborators
 - Visitors, Guests

Data and Resources

 - BHM: R-Codes / Packages
 - RealWorld Instrumental
 - RealWorld Proxies
 - Model Pseudo-Proxies
 - Noise models
 - Field-reconstructions

Publications/Talks

 - Reviewed
 - Reports
 - Talks
 - Bibliography

Education/Outreach

 - BHM-Concept
 - Slides
 - Podcasts
 - Teachers Materials
    "Little Ice Age":
   * Proxy Records
   * Math/Stats

 

Project Description

• Motivation and Goals:
  A primary concern of current climate research is how regional weather and climate is influenced by changes at the hemispheric and global level. Because climate in any particular region is often the result of multiple factors that operate at different time scales, long time series of data are required to identify and separate the contributions from these various processes. Because the instrumental record is limited in time, one has to use other kinds of measurements that act as proxies for direct meteorological measurements.
• Project Outline::
  These so-called paleoclimate proxy records, however, contain a host of uncertainties both with regard to their systematic representation of one or more climatic variables across multiple time scales (e.g., temperature, precipitation, cloudiness, soil moisture) and sometimes with regard to the actual timing of the variations. Current paleoclimate reconstruction methods are not designed to deal with such uncertainties but rely on a set of stationarity assumptions. Furthermore, the inherent uncertainty involved in the reconstructions themselves is either ignored or it is estimated after the fact from the unexplained variance during calibration. Little attention is paid to the fact that proxy records are single realizations of a noisy process which would more appropriately be represented by a family of possible outcomes.
  ->read on<-

Three Data Sources

• Tree-Ring Data
  Idealized view of high-resolution proxy record with good interannual to multi-decadal signal retention. Sensitivities - depending on location - to temperature, moisture availability, during growing season, incl. sometimes some memory ...
• Borehole Temperature Profiles
  Lowest resolution record providing the only direct measurement of past temperatures. Recovery of long-term change signal provides good estimate of mean change over centuries (very large and changing error for "timing").
• Pollen Records
  High-resolution