Activity Number:
|
259
- SPEED: Environmetrics: Spatio-Temporal and Other Models
|
Type:
|
Contributed
|
Date/Time:
|
Monday, July 30, 2018 : 3:05 PM to 3:50 PM
|
Sponsor:
|
Section on Statistics and the Environment
|
Abstract #332623
|
|
Title:
|
Spline Smoothing in Dendrochronology
|
Author(s):
|
Nicholas Bussberg* and Justin Maxwell and Scott Robeson and Chunfeng Huang
|
Companies:
|
Indiana University and Indiana University and Indiana University and Indiana University
|
Keywords:
|
smoothing parameter selection;
frequency response curves;
Fourier transformation;
tree-ring data
|
Abstract:
|
The cubic smoothing spline has been a popular data pre-processing step in the detrending of tree-ring data since the 1980's. We show this prevailing procedure does not consider end-point adjustments and incorrectly chooses the smoothing parameter. These issues lead to an inaccurate spline fit and interpretation. In this research, we provide the corrected form for the spline with the proper interpretation. Using simulations and tree-ring data, we show how the spline fit differs between the existing approach and our approach. The results indicate that a tree core of 250-year length can have as much as a 15% difference in the target frequency. We also find that tree-core series with more annual variation and more curvature lead to larger differences between the two approaches.
|
Authors who are presenting talks have a * after their name.