Redundancy analysis and the evolutionary learning algorithm as complementary processing tools for dendrochronological data

Abstract

To  process dendrochronological data sets, mathematical techniques that can  handle complexity are needed. Two methods from the field of numerial ecology  are introduced in tree ring analysis: redundancy analysis (an eigenvector  method) and the evolutionary learning algorithm (a machine learning tool).  Both methods show to be appropriate for a stringent test case. Redundancy  analysis explains variance in tree ring data by environmental date revealing  main trends. The evolutionary learning algorithm can be applied to look for  unexpected strong environmental signals possibly departing from main  trends.