NoAnalogueError

Potential Authors:

• Simon Goring
• J Sakari Salonen
• Miska Luoto
• Jack Williams
• Jing Gao

Testing pollen-based climate reconstructions in non-analogue space.

Much of this work is being laid out using two (or three) relatively fast methods of climate reconstruction, MAT, WA and WAPLS. This will allow us to develop some simple plots and outline ideas for discussion.

To Do list should go to the issues page.

Brief Paper Outline

Introduction

• What are no analogues? How have they been defined?
• What methods are used for pollen-based climate reconstruction?
• How can understanding how these methods work help us?

Methods

• Outline of each of the methods used
  • MAT
  • WA (with monotonic deshrinking)
  • WAPLS
  • randomForest
  • BRTs
• Outline of the bootstrap method
• Outline of Bias/Variance analysis

Results

• General description of error across methods
  • Related to analogue-distance
  • Related to spatial distribution
• Method specific notes
• Spatial