On October 25, 2017, I discussed my research on using paleoecological data (e.g., pollen) for paleoclimatic field reconstructions. Specifically, I explained how transfer functions can be used to infer past climate conditions from fossil pollen using modern pollen assemblages and climate condition relationships.
Modern pollen, climate, and location data came from the ‘analogue’ package. The data was compiled by Whitmore et al. 2005 (see presentation for full reference) for the North American Modern Pollen Database. The fossil pollen data was accessed through the Neotoma Paleoecological Database. Modern pollen data can also be accessed through Neotoma, of which both the modern and fossil can be accessed in R through the ‘neotoma’ package. You can also use better climate datasets, such as PRISM or WorldClim. To make things easier for the presentation, I used a wrapper for the transfer function, which is the randomTF function and uses the Modern Analogue Technique (MAT) from the ‘palaeoSig’ package. In essence, this allows for the matching of fossil pollen samples to their closest modern analogue (the ones that are the least dissimilar). In the clim_notebook, the ten closest analgoues were selected to do the reconstruction.
In clim_notebookknit1, several other packages are also highlighted (in addition to the use of R Markdown), such as leaflet (allowing for interactive mapping), analogue::Stratiplot (plotting of pollen profiles), dplyr, reshape2, ggplot, analogue, and palaeoSig.
Some specifics on MAT for pollen-based paleoclimate reconstruction and references can be found in my presentation, “Pollen-Based Paleo-Temperature Reconstructions.”
The additional script that you will need to run the code (the run_tf function, which is the wrapper for the ‘randomTF’ function from the ‘palaeoSig’ package) can be found here.
Finally, the knitted version of the clim_notebook R Markdown can be found here.