Article ID Journal Published Year Pages File Type
85633 Dendrochronologia 2013 7 Pages PDF
Abstract

To understand the effect of climate on tree-ring features, such as width or density, tree-ring data have to be calibrated against instrumental records. The high degree of multicollinearity among monthly time series of climate data violates the assumption of independent predictor variables in ordinary least squares regression. Bootstrapped confidence intervals of parameter estimates via regression against the principle components of the predictor variables are a possible solution to that problem. Package bootRes for R implements a flexible interface for bootstrapped response and correlation function analysis and tackles some shortcomings of currently available software. Given the increasing popularity of the free R software for statistical analysis, bootRes should facilitate both using R as a computational environment among tree-ring scientists and implementing new approaches to dendroclimatic calibration.

Related Topics
Physical Sciences and Engineering Earth and Planetary Sciences Atmospheric Science
Authors
, ,