Nonlinear Approaches to an Improved Understanding of West Antarctic Paleoclimate
Artificial neural networks (ANN) provide an opportunity to improve our understanding of West Antarctic climate by providing new nonlinear tools for prediction and analysis. ANN-based prediction supports development of a complete automatic weather station (AWS) record for the last twenty years, filli...
| Main Author: | |
|---|---|
| Other Authors: | |
| Format: | Text |
| Language: | English |
| Published: |
2001
|
| Subjects: | |
| Online Access: | http://citeseerx.ist.psu.edu/viewdoc/summary?doi=10.1.1.87.2617 http://www.geosc.psu.edu/~dbr/GradSchool_PSU/Dissertation Proposal.pdf |
| Summary: | Artificial neural networks (ANN) provide an opportunity to improve our understanding of West Antarctic climate by providing new nonlinear tools for prediction and analysis. ANN-based prediction supports development of a complete automatic weather station (AWS) record for the last twenty years, filling in gaps due to equipment failure and extending records to pre-date station installation. Traditional linear methods for extracting variance modes, principal component analysis (PCA) and canonical correlation analysis (CCA), are expanded into the nonlinear domain with new ANN-based methods. Nonlinear PCA is used to explore nonlinear behavior within West Antarctic glaciochemical and atmospheric circulation data sets. Nonlinear CCA is used to explore possible relationships between the glaciochemistry and the regional atmospheric circulation. All of these tasks are pursued with the intent of improving our understanding of modern West Antarctic climate and our interpretations of ice-core-based proxies for paleoclimate. Furthermore, this work seeks a suite of new nonlinear tools suited to understanding the nonlinear climate system of this region. 1. |
|---|